Yeast Functional Genomics and Proteomics: Methods and Protocols (Methods in Molecular Biology, 548) 1934115711, 9781934115718

Given the popularity and utility of Saccharomyces cerevisiae, yeast-based functional genomics and proteomics technologie

157 111

English Pages 317 [303] Year 2009

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Stagljar_FM
Stagljar_Ch01
Stagljar_Ch02
Stagljar_Ch03
Stagljar_Ch04
Stagljar_Ch05
Stagljar_Ch06
Stagljar_Ch07
Stagljar_Ch08
Stagljar_Ch09
Stagljar_Ch10
Stagljar_Ch11
Stagljar_Ch12
Stagljar_Ch13
Stagljar_Ch14
Stagljar_Ch15
Stagljar_Index
Stagljar_Color Plates
Recommend Papers

Yeast Functional Genomics and Proteomics: Methods and Protocols (Methods in Molecular Biology, 548)
 1934115711, 9781934115718

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

METHODS

IN

MOLECULAR BIOLOGY™

Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK

For other titles published in this series, go to www.springer.com/series/7651

Yeast Functional Genomics and Proteomics Methods and Protocols

Edited by

Igor Stagljar Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada

Editor Igor Stagljar, Ph.D. Department of Biochemistry & Department of Molecular Genetics Terrence Donnelly Centre for Cellular and Biomolecular Research University of Toronto Toronto, ON Canada

ISSN: 1064-3745 e-ISSN: 1940-6029 ISBN: 978-1-934115-71-8 e-ISBN: 978-1-59745-540-4 DOI: 10.1007/978-1-59745-540-4 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2008944036 © Humana Press, a part of Springer Science+Business Media, LLC 2009 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Cover illustration: Taken from Chapter 5, Figure 6B. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Dedication This book is dedicated to my parents Sonja and Mirko Stagljar for their great guidance and support in my life

Preface Often defined as “the unicellular human” and “everybody’s favorite fungus,” the baker’s yeast Saccharomyces cerevisiae has long been considered one of the most highly studied model organisms in the study of basic cellular processes. Along with this notion, yeastbased functional genomics and proteomics technologies, developed over the past decade, have contributed greatly to our understanding of bacterial, yeast, fly, worm, and human gene functions. More than 1,000 different papers and hundreds of reviews dealing with functional genomics and proteomics in yeast have appeared, but no comprehensive yeastbased functional genomics and proteomics textbook has yet been written and published. This book aims to be the standard textbook in the field of yeast-based functional genomics and proteomics and should serve as a stand-alone protocols handbook suitable for daily use in research laboratories. It includes recent advanced protocols in addition to major basic yeast-based functional genomics and proteomics techniques. In this way, both yeast researchers and those who wish to use yeast as a model system for functional genomics and proteomics will find this book useful. Chapter “Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast” serves as an introduction in how to use DNA microarrays to detect copy number variations in yeast. Chapter “Identification of Transcription Factor Targets by Phenotypic Activation and Microarray Expression Profiling in Yeast” describes in detail a methodology showing how overexpression of all yeast transcription factors combined with DNA microarray expression profiling and data analysis can be used to identify DNA-binding sequences for transcription factors. Chapter “SGAM: An Array-Based Approach for High-Resolution Genetic Mapping in Saccharomyces cerevisiae” contains state-of-the-art protocols for one of the best-known yeast functional genomics techniques, the synthetic genetic array (SGA) analysis, and focuses on a specific SGA application for high-resolution genetic mapping, referred to as SGA mapping (SGAM). Chapter “Reporter-Based Synthetic Genetic Array Analysis: A Functional Genomics Approach for Investigating the Cell Cycle in Saccharomyces cerevisiae” describes a modification of the SGA, termed reporterbased SGA (R-SGA) analysis, and its application in studying the expression of all yeast genes under a particular condition. Chapter “The Fidgety Yeast: Focus on HighResolution Live Yeast Cell Microscopy” gives an excellent insight into experimental strategies for live yeast cell imaging, geared towards imaging-based large-scale screens, whereas Chapter “A Genomic Approach to Yeast Chronological Aging” describes a novel functional genomics approach for quantitatively measuring the yeast chronological life span. Chapters “Chemogenomic Approaches to Elucidation of Gene Function and Genetic Pathways” and “Identification of Inhibitors of Chromatin Modifying Enzymes Using the Yeast Phenotypic Screens” contain series of protocols that were essentially invented to study drug action in yeast and thus set up a foundation for yeast-based chemical genomics approaches. Chapter “Exploiting Yeast Genetics to Inform Therapeutic Strategies for Huntington’s Disease” shows a perfect example of how yeast functional genomics approaches can efficiently be used to study a devastating human neurodegenerative disorder, Huntington’s disease. Chapter “Global Proteomic Analysis of Saccharomyces cerevisiae

vii

viii

Preface

Identifies Molecular Pathways of Histone Modifications” describes a proteomics method, the global proteomic analysis in S. cerevisiae (GPS), for the global analysis of the molecular machinery required for proper histone modifications. Chapters “Systematic Characterization of the Protein Interaction Network and Protein Complexes in Saccharomyces cerevisiae Using Tandem Affinity Purification and Mass Spectrometry,” “Protein Microarrays,” “Analysis of Protein–Protein Interactions Using Array-Based Yeast Two-Hybrid Screens,” and “Analysis of Membrane Protein Complexes Using the Split-Ubiquitin Membrane Yeast Two-Hybrid System” contain a collection of protocols for studying protein complexes and protein–protein interactions such as tandem affinity purification (TAP) linked to mass spectrometry, protein microarrays, the array-based yeast two-hybrid approach, and membrane yeast two-hybrid (MYTH) system. Protocols described in the last chapter aim to describe how computational analyses help us to understand the yeast proteome. Finally, I wish to take this opportunity to thank all authors for their great commitment, cooperation, and contributions that made my first editing job easier. I also wish to express my sincere thanks to Dr. John M. Walker for providing guidance on how to generate this book. Toronto, ON

Igor Stagljar

Contents Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2

3

4

5 6 7

8

9

10

11

12

Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Barry Dion and Grant W. Brown Identification of Transcription Factor Targets by Phenotypic Activation and Microarray Expression Profiling in Yeast . . . . . . . . . . . . . . . . . . . . Gordon Chua SGAM: An Array-Based Approach for High-Resolution Genetic Mapping in Saccharomyces cerevisiae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Costanzo and Charles Boone Reporter-Based Synthetic Genetic Array Analysis: A Functional Genomics Approach for Investigating the Cell Cycle in Saccharomyces cerevisiae. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Holly E. Sassi, Nazareth Bastajian, Pinay Kainth, and Brenda J. Andrews The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy . . . . . . Heimo Wolinski, Klaus Natter, and Sepp D. Kohlwein A Genomic Approach to Yeast Chronological Aging . . . . . . . . . . . . . . . . . . . . . . . Christopher R. Burtner, Christopher J. Murakami, and Matt Kaeberlein Chemogenomic Approaches to Elucidation of Gene Function and Genetic Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sarah E. Pierce, Ronald W. Davis, Corey Nislow, and Guri Giaever Identification of Inhibitors of Chromatin Modifying Enzymes Using the Yeast Phenotypic Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benjamin Newcomb and Antonio Bedalov Exploiting Yeast Genetics to Inform Therapeutic Strategies for Huntington’s Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Flaviano Giorgini and Paul J. Muchowski Global Proteomic Analysis of Saccharomyces cerevisiae Identifies Molecular Pathways of Histone Modifications . . . . . . . . . . . . . . . . . . . . Jessica Jackson and Ali Shilatifard Systematic Characterization of the Protein Interaction Network and Protein Complexes in Saccharomyces cerevisiae Using Tandem Affinity Purification and Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mohan Babu, Nevan J. Krogan, Donald E. Awrey, Andrew Emili, and Jack F. Greenblatt Protein Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joseph Fasolo and Michael Snyder

ix

vii xi 1

19

37

55

75 101

115

145

161

175

187

209

x

Contents

13

Analysis of Protein–Protein Interactions Using Array-Based Yeast Two-Hybrid Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Seesandra V. Rajagopala and Peter Uetz 14 Analysis of Membrane Protein Complexes Using the Split-Ubiquitin Membrane Yeast Two-Hybrid (MYTH) System . . . . . . . . . . . . . . . . . . . . . . . . . . Saranya Kittanakom, Matthew Chuk, Victoria Wong, Jamie Snyder, Dawn Edmonds, Apostolos Lydakis, Zhaolei Zhang, Daniel Auerbach, and Igor Stagljar 15 Computational Analysis of the Yeast Proteome: Understanding and Exploiting Functional Specificity in Genomic Data . . . . . . . . . . . . . . . . . . . . . . . . Curtis Huttenhower, Chad L. Myers, Matthew A. Hibbs, and Olga G. Troyanskaya Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

223

247

273

295

Contributors BRENDA J. ANDREWS • Department of Molecular Genetics, Banting and Best Department of Medical Research, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada DANIEL AUERBACH • DUALSYSTEMS Biotech Inc., Zurich, Switzerland DONAL E. AWREY • Campbell Family Institute for Breast Cancer Research Therapeutics, Toronto, ON, Canada MOHAN BABU • Banting and Best Department of Medical Research, University of Toronto, Terrence Donnelly Center for Cellular and Biomolecular Research, Toronto, ON, Canada NAZARETH BASTAJIAN • Department of Medical Genetics and Microbiology, Banting and Best Department of Medical Research, University of Toronto, Toronto, ON, Canada ANTONIO BEDALOV • Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA CHARLIE BOONE • Department of Molecular Genetics, Banting and Best Department of Medical Research, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada GRANT W. BROWN • Department of Biochemistry and Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada CHRISTOPHER R. BURTNER • Department of Biochemistry, University of Washington, Seattle, WA, USA GORDON CHUA • Institute for Biocomplexity and Informatics, Department of Biological Sciences, University of Calgary, Calgary, AB, Canada MATTHEW CHUK • Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada MICHAEL COSTANZO • Banting and Best Department of Medical Research, University of Toronto, Toronto, ON, Canada RON W. DAVIS • Stanford Genome Technology Center, Stanford University, Palo Alto, CA, USA BARRY DION • Department of Biochemistry, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada DAWN EDMONDS • Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada ANDREW EMILI • Department of Molecular Genetics, Banting and Best Department of Medical Research, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada

xi

xii

Contributors

JOSEPH FASOLO • Department of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT, USA GURI GIAEVER • Department of Pharmaceutical Sciences, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada FLAVIANO GIORGINI • Department of Genetics, University of Leicester, Leicester, UK JACK F. GREENBLATT • Department of Molecular Genetics, Banting and Best Department of Medical Research, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada MATTHEW A. HIBBS • Lewis-Sigler Institute for Integrative Genomics, Carl Icahn Laboratory, Princeton University, Princeton, NJ, USA; Department of Computer Science, Princeton University, Princeton, NJ, USA CURTIS HUTTENHOWER • Lewis-Sigler Institute for Integrative Genomics, Carl Icahn Laboratory, Princeton University, Princeton, NJ, USA; Department of Computer Science, Princeton University, Princeton, NJ, USA JESSICA JACKSON • Department of Genetics, Washington University School of Medicine, St. Louis, MO, USA MATT KAEBERLEIN • Department of Pathology, University of Washington, Seattle, WA, USA PINAY KAINTH • Department of Medical Genetics and Microbiology, Banting and Best Department of Medical Research, University of Toronto, Toronto, ON, Canada SARANYA KITTANAKOM • Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada SEPP D. KOHLWEIN • Institute of Molecular Biosciences, University of Graz, Graz, Austria NEVAN J. KROGAN • Department of Cellular and Molecular Pharmacology, University of California, San Francisco, CA, USA APOSTOLOS LYDAKIS • Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada PAUL J. MUCHOWSKI • Departments of Biochemistry and Biophysics, and Neurology, Gladstone Institute of Neurological Disease, University of California, San Francisco, San Francisco, CA, USA CHRISTOPHER J. MURAKAMI • Department of Pathology, University of Washington, Seattle, WA, USA CHAD L. MYERS • Department of Computer Science, University of Minnesota, Minneapolis, MN, USA KLAUS NATTER • Institute of Molecular Biosciences, University of Graz, Graz, Austria BENJAMIN NEWCOMB • Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA COREY NISLOW • Department of Molecular Genetics, Banting and Best Department of Medical Research, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada SARAH E. PIERCE • Stanford Genome Technology Center, Stanford University, Palo Alto, CA, USA

Contributors

xiiiiixi

SEESANDRA V RAJAGOPALA • J. Craig Venter Institute, Rockville, MD, USA HOLLY E. SASSI • Department of Medical Genetics and Microbiology, Banting and Best Department of Medical Research, University of Toronto, Toronto, ON, Canada ALI SHILATIFARD • Stowers Institute for Medical Research, Kansas City, MO, USA JAMIE SNYDER • Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada MICHAEL SNYDER • Department of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT, USA IGOR STAGLJAR • Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada OLGA G. TROYANSKAYA • Lewis-Sigler Institute for Integrative Genomics, Carl Icahn Laboratory, Princeton University, Princeton, NJ, USA; Department of Computer Science, Princeton University, Princeton, NJ, USA PETER UETZ • J. Craig Venter Institute, Rockville, MD, USA HEIMO WOLINSKI • Institute of Molecular Biosciences, University of Graz, Graz, Austria VICTORIA WONG • Department of Biochemistry & Department of Molecular Genetics, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada ZHAOLEI ZHANG • Department of Molecular Genetics, Banting and Best Department of Medical Research, Terrence Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, ON, Canada

Chapter 1 Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast Barry Dion and Grant W. Brown Summary Chromosomal aberrations resulting in aneuploidies have been implicated in the development of most cancers and numerous other genetic disorders. Aneuploidies are a key feature of genomic instability, so classification of these copy number changes will be important in understanding how rearrangements arise and how ongoing instability is maintained. Traditional methods for detecting copy number changes have relatively poor resolution, making accurate detection of breakpoints impossible. The advent of microarray technology and its advance over the years has improved the ability to detect aneuploidies with greater accuracy. Mammalian comparative genome hybridization on microarrays (array-CGH) has been applied to the study of many carcinomas, identifying common copy number changes in key regions including known oncogenes. However, the large size of mammalian genomes has made it impractical to perform whole genome CGH at high resolution. Yeast has been established as a useful model for studying pathways relevant to oncogenesis, particularly those that maintain the integrity of the genome. Given the smaller size of the yeast genome, oligonucleotide tiling arrays have been developed that allow for nucleotide resolution of the whole genome on a single chip. Here we describe in detail how to use these arrays to detect copy number variations in yeast. This method will be useful in many different studies, but particularly in monitoring and cataloguing the changes resulting from genetic instability. Key words: Comparative genome hybridization, Tiling microarrays, Aneuploidy, Array-CGH, Whole genome amplification

1. Introduction Chromosomal abnormalities such as amplifications, deletions, and translocations resulting in DNA copy alterations have been implicated in the development of most cancers and a number of other human genetic disorders. Specifically, amplifications of

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_1, © Humana Press, a part of Springer Science + Business Media, LLC 2009

1

2

Dion and Brown

oncogenes, deletions of tumor suppressors, and oncogenic fusions have all been identified in human cancers (1) and can contribute to tumorigenesis. Other chromosomal aberrations, such as duplications and deletions can result in genetic diseases such as Down syndrome and Cri du Chat syndrome (2, 3). Given the obvious importance of genomic instability in the development of many pathologies it is clear that the methods to detect and map aneuploidies are of great utility in determining the mechanisms by which these aneuploidies arise. In the past, comparative genome hybridization (CGH) has been applied to metaphase chromosome spreads to assess DNA copy number changes in mammalian cells. However, this method was limited by its low resolution, between 10 and 20 Mb (4, 5). In its modern form, CGH is performed on microarrays which involves the labeling of reference and sample DNA with different fluorochromes. The DNA is then competitively hybridized to DNA probes representing the whole or partial genome of interest. The resultant signal intensity ratio between the two samples corresponds to the copy number imbalance. Array-CGH, first performed in 1997, improved the resolution level to 75–130 kb, compared to previous CGH methods (6). Advances in array technologies since then have allowed a vast improvement in the resolution capabilities of array-CGH. Since CGH requires a net change in DNA content to detect aneuploidies, it is unable to detect reciprocal translocations. Despite this limitation, array-CGH is seeing increased use as a means of comparing genomes for the purpose of identifying alterations resulting in copy number changes. In mammalian cells, array-CGH has been effective in detecting aneuploidies that are characteristic of many human cancers and genetic diseases. For example, studies of fallopian tube carcinoma have revealed a large number of copy number changes (7, 8). The improved resolution of array-CGH has been applied to further identify and refine regions of copy number alterations quantitatively while mapping the aberrations directly to the human genome sequence at a resolution of ~1.4 Mb (9). Mapping of these genetic changes revealed recurrent amplifications in a number of known oncogenes (9). Colorectal cancers have also been surveyed by array-CGH at 1–2 Mb resolution, identifying high frequency losses and gains previously identified by metaphase CGH, as well as additional recurrent aberrations that were not previously identified (10). More recently array-CGH studies in breast cancers have identified common copy number changes to a resolution of 100 kb on chromosome 8 (11). These are only a few of the many type of cancers for which array-CGH has been able to identify changes in DNA copy number which may play an important role in tumorigenesis. Array-CGH has allowed researchers and clinicians to further define the characteristics of various cancers, identifying previously unidentified aberrations,

Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast

3

to complement those alterations already known. This knowledge can be used to further our understanding of the cause and progression of cancers, in addition to providing clinicians with an efficient method for diagnosis. Despite the improvement in CGH using microarrays, mammalian array-CGH is still hindered by relatively poor resolution. To date only sub-megabase resolution microarrays have been developed that span the whole human genome (12). Although breakpoint mapping information obtained from low resolution arrays can be used to create specific tiling oligonucleotide arrays that allow for CGH at high resolution (13), the large size of mammalian genomes remains a challenge in high resolution array-CGH. The relatively small genome of Saccharomyces cerevisiae allows array-CGH to be performed at a much higher resolution compared to mammalian studies to date. State of the art tiling microarrays span the entire genome on a single chip at 4 bp resolution (14, 15). A variety of studies have employed the use of these yeast whole-genome tiling arrays to monitor copy number changes. Adaptive rearrangements resulting from nutrient limitation during experimental evolutions have been mapped at single gene resolution (16). Array-CGH has also been used to monitor the efficiency of DNA replication origins and the timing of the initiation of DNA replication (17). CGH on microarrays has also been used to compare different yeast strains and species, demonstrating its usefulness in species determination and differentiation of strains within a species (18). Clearly array-CGH can be applied to a wide variety of questions addressing patterns and rates of changes during genome evolution (19). S. cerevisiae has seen extensive use as a model in the analysis of genomic instability, motivating and informing further studies in mammalian systems. Examples of this include the discovery of cell cycle checkpoints (20, 21), checkpoint mediators such as Rad9 and Mrc1 (22, 23), and the gross chromosomal rearrangement studies of the Kolodner lab, which helped to define the genetic basis for genomic instability (24, 25). Further research is needed to advance our limited understanding of how genome rearrangements arise, the pathways that suppress them, and whether defects in these pathways result in ongoing instability that is present in many cancers (24). Understanding genome rearrangement is key to understanding the causes and effects of genomic instability and requires high resolution techniques to identify aberrations and to map their breakpoints accurately. Array technology has not yet produced affordable methods by which to study genomic aberrations at high resolution in mammalian cells. The availability of these high resolution technologies for yeast provides a convenient model system by which to elucidate the molecular mechanisms behind genome instability and uncontrolled cell growth. Given

4

Dion and Brown

that many cancers arise as a result of a loss of genome stability, these studies in yeast can be used to direct studies in mammalian cells more efficiently, providing further mechanistic insight into the causes of cancers. In this chapter we outline a procedure to detect aneuploidies (changes in DNA copy number), such as amplifications and deletions, at high resolution in yeast. The basic protocol includes DNA isolation, amplification (where necessary), DNA fragmentation and labeling, array hybridization, and data analysis (Fig. 1). The array used for detection of these aneuploidies is a single oligonucleotide tiling array containing 6.5 million probes interrogating both strands of S. cerevisiae genomic sequence with 25-mer probes tiled at eight nucleotide intervals (26). This array allows for high resolution mapping of breakpoints, and with more sophisticated data analysis can allow for the detection of single nucleotide polymorphisms (14).

Fig. 1. Schematic of the array-CGH method for detection of aneuploidies.

Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast

5

2. Materials 2.1. DNA Isolation 2.1.1. Yeast Genomic DNA Preparation

1. Genomic DNA Midi-prep Kit (Qiagen Genomic-tip 100/G Cat# 10243). 2. Genomic DNA buffer set (Qiagen Cat# 19060). 3. YPD (see Subheading 2.7). 4. 125-ml Erlenmeyer flasks. 5. RNase A (see Subheading 2.8). 6. Zymolyase. 7. Proteinase K (see Subheading 2.8). 8. 100% Isopropanol. 9. 70% Ethanol. 10. 50-ml Conical centrifuge tubes. 11. 15-ml Conical centrifuge tubes. 12. Polycarbonate round bottom centrifuge tubes. 13. 1.5-ml microfuge tubes. 14. 10 mM Tris–HCl (see Subheading 2.8).

2.1.2. Yeast Genomic DNA Mini-Prep

1. YeaStar genomic DNA kit (Zymo Research Cat# D2002). 2. Culture tubes. 3. YPD (see Subheading 2.7). 4. 1.5-ml microfuge tubes. 5. 10 mM Tris–HCl (see Subheading 2.8). 6. Chloroform.

2.1.3. DNA Quantification

1. Quant-iT dsDNA HS assay kit (Invitrogen Cat# Q32851). 2. Qubit assay tubes (Invitrogen Cat# Q33856). 3. Qubit fluorometer (Invitrogen Cat# Q32857). 4. DNA to be quantified.

2.2. Whole Genome Amplification

1. GenomePlex complete whole genome amplification kit (Sigma Cat# WGA2). 2. Thin-walled PCR tubes. 3. DNA to be amplified.

2.3. Purification of Amplified DNA

1. QIAquick PCR purification kit (Qiagen Cat# 28104). 2. Microfuge tubes. 3. Amplified DNA.

6

Dion and Brown

2.4. DNA Fragmentation

1. 10× One-phor-all buffer PLUS (GE Healthcare Amersham Cat# 27-091-02). 2. Cobalt chloride (CoCl2) (see Subheading 2.8). 3. Deoxyribonuclease I, amplification grade (Invitrogen Cat# 18068-015). 4. Thin-walled PCR tubes. 5. Agarose. 6. 20× Lithium boric (LB) acid ultralow-conductive medium for DNA electrophoresis (Faster Better Media LLC Cat# LB20). 7. 5× LB loading medium (Faster Better Media LLC Cat# LB5N). 8. SYBR Green. 9. TAE (see Subheading 2.8). 10. 10 bp DNA ladder molecular weight standards (Invitrogen Cat# 10821-015). 11. Sterile ddH2O. 12. DNA to be fragmented.

2.5. DNA Labeling

1. Biotin-N6-ddATP (Enzo Life Sciences Cat# 42809). 2. Terminal deoxynucleotidyl transferase (MBI Fermentas Cat# EP0161). 3. Fragmented DNA. 4. Thin-walled PCR tubes.

2.6. Array Hybridization and Washing

1. Hybridization buffer (see Subheading 2.8). 2. Sterile dH2O (filtered). 3. Bovine serum albumin (BSA) (see Subheading 2.8). 4. Herring testes carrier DNA, denatured (Clonetech Cat# 50277). 5. b213 control oligonucleotide: Biotin-CTG AAC GGT AGC ATC TTG AC 3¢ 6. MES stain buffer (see Subheading 2.8). 7. Streptavidin, R-phycoerythrin conjugate (SAPE) (Invitrogen Cat# S866). 8. Biotinylated anti-streptavidin antibody (Vector Laboratories Cat# BA-500). 9. Normal goat IgG (see Subheading 2.8) (Sigma Cat# I 5256). 10. 20× SSPE (UltraPure) (Invitrogen Cat# 15591). 11. Wash A: Non-stringent wash buffer (see Subheading 2.8). 12. Wash B: Stringent wash buffer (see Subheading 2.8). 13. Microtube tough-spots (Mandel Cat# US-9185-050X).

Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast

7

14. Compressed gas duster. 15. Lens paper. 16. 95% Ethanol. 17. Microfuge tubes. 18. Biotin labeled DNA. 19. S. cerevisiae Tiling Array (Affymetrix Cat# 520055). 2.7. Media

1. YPD (1 l liquid): 10 g yeast extract, 20 g bio-tryptone, 20 g d-glucose. Resuspend all ingredients in 1 l ddH2O, dissolve by stirring. Autoclave at 121°C, 15 psi for 30 min.

2.8. Solutions

1. 10 mM Tris–HCl: Dissolve 1.211 g Tris base into 800 ml ddH2O. Stir while adjusting pH to 8.5 with concentrated HCl. Adjust volume to 1 l with ddH2O and autoclave at 121°C, 15 psi for 30 min. 2. 50× TAE (27): Dissolve 242 g Tris base in 650 ml ddH2O. Add 57.1 ml of glacial acetic acid and 100 ml 0.5 M EDTA (pH 8.0). Stir solutions with magnetic stir and adjust volume to 1 l with ddH2O. The working solution is 1× (40 mM Trisacetate, 1 mM EDTA). 3. 12× MES stock: Dissolve 7.74 g MES free acid monohydrate and 21.26 g MES sodium salt in 80 ml ddH2O. Stir with magnetic stirrer and adjust pH to between 6.5 and 6.7. Filter through a 0.22-mm polyethersulfone filter. Store at 4°C shielded from light. If solution turns yellow, discard and make a new batch. 4. 2× Hybridization buffer: Mix 8.3 ml 12× MES stock, 17.7 ml 5 M NaCl, 4 ml 0.5 M EDTA, 0.1 ml 10% Tween 20, 19.9 ml ddH2O together. Filter through a 0.22-mm polyethersulfone filter. Store at 4°C shielded from light. If solution turns yellow, discard and make a new batch. 5. 1× Hybridization buffer: Dilute 2× hybridization buffer with ddH2O by half to a final concentration of 100 mM MES, 1 M NaCl, 0.05% Tween 20. Filter through a 0.22-mm polyethersulfone filter. Store at 4°C shielded from light. If solution turns yellow discard, and make a new batch. 6. Wash A: Non-stringent wash buffer: Mix 300 ml 20× SSPE, and 1 ml 10% Tween 20 with 699 ml ddH2O. 7. Wash B: Stringent wash buffer: Mix 83.3 ml 12× MES stock, 5.2 ml 5 M NaCl and 1 ml 10% Tween 20 with 910.5 ml ddH2O. Protect from light with aluminum foil and store at 4°C. If solution turns yellow discard, and make a new batch. 8. 30 mM CoCl2: Dissolve 35.69 mg cobalt chloride hexahydrate in 5 ml ddH2O. Filter through a 0.22-mm polyethersulfone filter and store at room temperature.

8

Dion and Brown

9. 20 mg/ml BSA: Dissolve 200 mg of bovine albumin (fraction V) into 10 ml of ddH2O by vortexing. Filter through a 0.22 mm polyethersulfone sterilizing filter and store at −20°C. 10. 30 mg/ml normal goat IgG: Reconstitute 30 mg normal goat IgG with 1 ml 150 mM NaCl solution. Aliquot into smaller portions and store at −20°C. Do not freeze and thaw repeatedly. 11. 5 M NaCl (27): Dissolve 292.2 g of sodium chloride in 800 ml ddH2O. Adjust volume to 1 l with ddH2O and autoclave at 121°C, 15 psi for 30 min. 12. 10% Tween 20: Mix 10 ml Tween 20 with 90 ml ddH2O, stir with magnetic stir bar. Filter through a 0.22-mm polyethersulfone filter and store at 4°C. 13. 0.5 M EDTA (pH 8.0) (27): Dissolve 186.1 g of disodium ethylenediaminetetra-acetate·2H2O with 800 ml ddH2O. Stir with magnetic stirrer while adjusting pH to 8.0 with NaOH. Adjust volume to 1 l with ddH2O and autoclave at 121°C, 15 psi for 30 min. 14. 20 mg/ml Proteinase K: Dissolve 250 mg of proteinase K (activity > 30 U/mg) into 2.5 ml ddH2O by vortexing gently. Aliquot into 100 ml fractions and store at −20°C. 15. 100 mg/ml RNase A: Dissolve 100 mg of DNase-free ribonuclease A (>60 K U/mg) in 1 ml ddH2O. Vortex and store at −20°C. 16. 3 M sodium acetate (pH 5.2) (27): Dissolve 102 g of sodium acetate·3H2O in 200 ml ddH2O. Adjust pH to 5.2 with glacial acetic acid. Adjust volume to 250 ml with ddH2O. Autoclave at 121°C, 15 psi for 30 min to sterilize.

3. Methods 3.1. Yeast Genomic DNA Midi-Prep

1. Inoculate 35 ml of YPD with a single yeast colony. 2. Grow overnight, shaking (180 rpm) at 30°C. 3. Measure density of cells in a spectrophotometer at 600 nm. 4. Harvest 7.0 × 109 cells (OD600 value of 1 equals 2 × 107 cells/ml) at 2,100 × g for 5 min. 5. Extract genomic DNA using Qiagen Genomic-tip 100/G kit (Cat# 10243) and Qiagen Genomic DNA Buffer set (Cat# 19060) (or with equivalent product) with the following modifications: (a) Incubate with zymolyase at 30°C for 1 h. (b) Incubate with proteinase K at 50°C for 1 h.

Comparative Genome Hybridization on Tiling Microarrays to Detect Aneuploidies in Yeast

9

(c) Prewarm elution buffer QF to 50°C. (d) Pellet precipitated DNA by centrifugation at 23,500 × g for 15 min for both the isopropanol and 70% ethanol steps. (e) Resuspend pelleted DNA in 400 ml of 10 mM Tris–HCl, thoroughly washing the sides of the centrifuge tube. (f) Transfer to 1.5-ml microfuge tube and dissolve overnight or at 55°C for 2 h. 6. Measure concentration of genomic DNA (see Subheading 3.8). 7. Concentrate genomic DNA if concentration is less than 0.403 mg/ml (see Subheading 3.10). 3.2. Yeast Genomic DNA Mini-Prep

1. Inoculate 5 ml of YPD with single colony (see Note 1). 2. Grow overnight at 30°C with rotation. 3. Extract genomic DNA using YeaStar Genomic DNA kit (Cat# D2002) Protocol I or with equivalent product with the following modifications: (a) Spin 5 × 107 cells at 2,700 × g for 2 min (see Note 1). (b) Incubate with R-zymolyase (provided in kit) for 60 min and vortex for 1 min. (c) Vortex YD lysis buffer with sample for 1 min at mediumlow setting. (d) Centrifuge at 17,900 × g for 2 min. (e) Elute in 60 ml of 10 mM Tris–HCl and centrifuge for 30 s. 4. Measure the concentration of genomic DNA (see Subheading 3.8). 5. Concentrate genomic DNA if concentration is less than 5 ng/ml (see Subheading 3.10).

3.3. Whole Genome Amplification

1. Amplify genomic DNA prepared with genomic DNA mini-prep kit (see Subheading 3.2) using GenomePlex complete whole genome amplification kit (Cat# WGA2) with the following modifications: (a) Prepare a 5 ng/ml DNA solution from extracted DNA. (b) Add 10 ml of 5 ng/ml DNA solution to 1 ml of 10× fragmentation buffer. 2. Measure amplified DNA concentration (see Subheading 3.8) (see Note2). 3. Purify final product before use (see Subheading 3.9) (see Note 3).

3.4. DNA Fragmentation

1. Prepare DNase master mix: 7.4 ml ddH2O, 1 ml 10× onephor-all buffer, 0.6 ml 30 mM CoCl2 and 1 ml 1 U/ml DNase I (Invitrogen, amplification grade). Mix solution with a pipette.

10

Dion and Brown

2. Prepare 0.403 mg/ml DNA (from Subheadings 2.1 or 2.3) solution with ddH2O to a total volume of 37.25 mg/ml in a thin-walled PCR tube. Mix solution with a pipette. 3. Add to each DNA sample: 4.5 ml 10× one-phor-all buffer, 2.25 ml 30 mM CoCl2 and 1.5 ml of the DNase master mix. Mix solution with a pipette. 4. Use a thermocycler to incubate for 4 min at 37°C, then 95°C for 10 min, then decrease the temperature to 4°C (see Note 4). 5. Check digestion on a thin (0.3 between experimental replicates are indicative that the differentially expressed genes are more likely due to TF overexpression rather than microarray noise (5). The differentially regulated genes are also examined for significant enrichment in Gene Ontology functional categories using FunSpec to identify putative function and targets of TFs (6). Finally, the promoter regions of differentially expressed genes are subjected

Identification of Transcription Factor Targets by Phenotypic Activation Empty vector

21

Gal1-10-TF

Isolate mRNA

Reverse transcription and label with Cy3 and Cy5 dyes

Mix cDNA and hybridize on ORF microarray

Scan microarray, quantitate and normalize data

TF overexpression profile Differential gene expression: motif searching and functional enrichment Identify putative TF targets

Fig. 1. Flow chart describing the procedure involved in the microarray profiling of transcription factor overexpression strains and data analysis to identify putative targets.

to a motif-finding algorithm called RankMotif to identify putative TF-binding sites (3).

2. Materials 2.1. Cell Culture

1. Synthetic uracil dropout medium with 2% raffinose (SR-Ura). 2. 40% Galactose. 3. 250-ml baffled culture flasks. 4. 50-ml Falcon tubes. 5. Liquid nitrogen. 6. Spectrophotometer (microplate reader).

22

Chua

2.2. Total RNA Isolation

1. AE buffer: 50 mM sodium acetate, pH 5.2, 10 mM ethylenediamine tetraacetic acid (EDTA), 1% sodium dodecyl sulfate (SDS), diethylpyrocarbonate-treated (DEPC) H2O. 2. Unbuffered liquefied acid phenol (pH = 4.5–5.5) (this is extremely toxic and corrosive and care should be taken not to receive exposure). Store at 4°C. 3. Phenol: chloroform: isoamyl alcohol (25:24:1) (this is extremely toxic and corrosive and care should be taken not to receive exposure). Store at 4°C. 4. Chloroform: isoamyl alcohol (24:1). Store at 4°C. 5. 3 M sodium acetate, pH 5.2, DEPC H2O. 6. 95–100% Ethanol. 7. 70% Ethanol. 8. Isopropanol. 9. 0.1% DEPC H2O. 10. Acid-washed 425–600 µm glass beads (Sigma). 11. Plasticware: 10-ml disposal pipettes, 15-ml Falcon tubes. 12. Multi-tube vortexer (VWR).

2.3. Poly-A mRNA Isolation

1. 2× Loading buffer: 40 mM Tris–HCl, pH 7.6; 1 M NaCl; 2 mM EDTA; 0.2% sodium lauryl sarosine (SLS); DEPC water. 2. Middle wash buffer: 20 mM Tris–HCl, pH 7.6; 150 mM NaCl; 1 mM EDTA; 0.1% SLS; DEPC water. 3. Elution buffer: 10 mM Tris–HCl, pH 7.6; 0.1 mM EDTA; DEPC water. 4. Oligo-dT cellulose (SIGMA): This can be recycled by washing the resin with 0.1 N NaOH, which will elute and/or degrade any attached RNA. Store at 4°C. 5. Poly-prep columns (Bio-Rad). 6. 95–100% Ethanol. 7. 0.1 N NaOH. 8. DEPC H2O. 9. Linear acrylamide (Ambion). Store at −20°C. 10. 3 M sodium acetate, pH 5.2, DEPC water. 11. Plasticware: 10-ml disposal pipettes, 6-ml and 50-ml Falcon tubes. 12. Nutator.

2.4. Reverse Transcription and Aminoallyl Labeling of mRNA

1. PCR purification columns and associated reagents (Qiagen). 2. T18-VN primer (1 mg/ml).

Identification of Transcription Factor Targets by Phenotypic Activation

23

3. Superscript II reverse transcriptase with 5× buffer and 0.1 M dithiothreitol (Invitrogen). 4. 10 mM dNTPs. 5. 1 mM aminoallyl-dUTP (Sigma): Dissolve 1 mg of AAdUTP in 191 ml DEPC water. Store at −20°C. 6. 1 N NaOH. 7. 0.5 M EDTA. 8. 1 M Tris–HCl, pH 7.6. 9. 80% Ethanol. 10. DEPC H2O. 11. 42 and 65°C water baths. 12. Speed-Vac. 2.5. Cy3/Cy5 Dye Coupling of cDNA

1. Anhydrous dimethyl sulfoxide (DMSO, Sigma). 2. 2× Bicarb buffer: (one bicarbonate capsule (Sigma), 25 ml water, 125 ml 37% HCl). Note: this buffer should be made fresh and is good only for 1 week. 3. Cy3 and Cy5 monofunctional dye packs (GE Healthcare). 4. 4 M hydroxylamine. 5. PCR-purification (Qiagen).

2.6. Microarray Hybridization

columns

and

associated

reagents

1. Quackenbush prehybridization solution: 5× SSC, 1% SDS, and 1% BSA. Store in 50-ml aliquots at −20°C. 2. Hybridization solution: 5× SSC, 25% formamide, 0.1% SDS. Store in 1-ml aliquots at −20°C. 3. Wash 1 solution: 1× SSC, 0.2% SDS. 4. Wash 2 solution: 0.1× SSC, 0.2% SDS. 5. Wash 3 solution: 0.1× SSC. 6. Microscope slide mailers. 7. Lifter cover slips. 8. Curved forceps. 9. UV crosslinker. 10. Speed-Vac. 11. Rack for microscope slides. 12. Hybridization oven (42°C). 13. Hybridization chambers. 14. 250-ml glass dishes. 15. Compressed air (e.g., Dust-off).

24

Chua

2.7. Acquisition, Processing, and Analysis of Microarray Data

1. Laser microarray scanner with integrated software for quantitation and normalization of data (e.g., Axon GenePix 4200A scanner with GenePix Pro/Acuity software – Molecular Devices).

3. Methods 3.1. Cell Culture

1. For each experiment, prepare two overnight cultures of isogenic strains, one containing a 2-µm vector with a GAL1/10driven TF gene and the other containing an empty vector (pEGH, URA3+). Inoculate several medium-size colonies (~10) from an agar plate into 100 ml SR-Ura medium in a 250-ml baffled culture flask. In addition, set up a second diluted culture (1/5–1/10×) to ensure that one of the overnight cultures will be in the appropriate cell density range the following day. Shake cultures at 30°C overnight. 2. For each strain, select an overnight culture in mid-log phase (20 mg. Store mRNA in −80°C. 22. Recycling the oligo-dT cellulose and column: Cap the bottom of the columns and add 2 ml of 0.1 N NaOH to each column. Resuspend the oligo-dT cellulose in each column by pipetting, and transfer contents into a 50-ml Falcon tube. Add 1 ml of 0.1 N NaOH to the column, resuspend, and transfer remaining oligo-dT cellulose into the Falcon tube. It is important to remove as much residual oligo-dT cellulose from the column to prevent blockage in future mRNA isolations. Wash oligo-dT cellulose three times with DEPC water. After the final wash, add 50 ml DEPC water and store at 4°C. Rinse the columns twice with DEPC water before storage. 3.4. Reverse Transcription and Aminoallyl Labeling of cDNA

1. Each experiment will be performed twice with fluor reversal (control sample–Cy3 vs. experimental sample–Cy5, and experimental sample–Cy3 vs. control sample–Cy5). As a result, each mRNA sample will be reverse-transcribed twice. Aliquot equal amounts (1–2 mg) of control and experimental mRNA samples into microcentrifuge tubes. Dry down samples in a Speed-Vac (medium-heat setting to prevent RNA hydrolysis). 2. Dissolve each sample in 10.5 ml DEPC water and 1 ml T18VN primer (1 mg/ml). 3. Denature the samples at 65°C for 5 min. 4. Incubate at 42°C for 5 min to anneal the T18VN primer. 5. While the samples are incubating at 42°C, prepare a master reaction mixture containing 4 ml 5× RT buffer, 2 ml 0.1 M DTT, 1 ml 10 mM dNTPs, 1 ml 1 mM aminoallyl-dUTP, and

28

Chua

0.5 ml superscript II enzyme per sample. Add 8.5 ml of reaction mixture to each sample. 6. Incubate 50 min at 42°C. 7. Add 10 ml of a 1:1 mixture 0.5 M EDTA: 1 N NaOH to each of the samples. 8. Incubate at 65°C for 20 min to hydrolyze the RNA. During this incubation, set up and label a QIAGEN purification column and microcentrifuge tube for each sample, and preheat some water at 65°C. 9. Add 10 ml 1 M Tris–HCl pH 7.6 to each sample. 10. To purify the cDNA samples and to remove Tris to prevent undesired coupling of the monofunctional NHS-ester Cydyes to free amine groups in solution, add 60 ml water to each sample to bring the reaction volume to 100 ml. 11. Add 500 ml Qiagen buffer PB to each sample, mix well by pipetting, and apply to the PCR purification columns. 12. Centrifuge the columns at 3,800 × g for 1 min at room temperature and discard the flow-through. 13. Add 600 ml 80% ethanol to each column, centrifuge at 3,800 × g for 1 min, and discard the flow-through. 14. Repeat the ethanol washes two more times. 15. Centrifuge at 16,000 × g for 1 min to dry the column. 16. Elute each sample from the column into a microcentrifuge tube twice with 40 ml 65°C water. 17. Dry the cDNA samples in Speed-Vac (high-heat setting). 18. The cDNA pellet can now be frozen at −80°C for at least several weeks prior to coupling. 3.5. Cy3/Cy5 Dye Coupling of cDNA

1. Resuspend each of the dried cDNA samples in 3.5 ml water. 2. Arrange each pair of cDNA samples of which one will be labeled with Cy3 and the other with Cy5. 3. Preparation of Cy3 and Cy5 dyes: Resuspend each Cy3 and Cy5 dye pack in 15 ml anhydrous DMSO by pipetting up and down 30–50 times, and vortexing for 1 min, prior to pulse centrifugation. Immediately recap and reseal the DMSO. Each dye pack can label 12–14 cDNA samples. 4. Dye coupling: Add 30 ml 2× Bicarb buffer to the Cy3 dye. Quickly aliquot 3.5 ml to each of the Cy3 cDNA samples. Repeat for Cy5 samples. 5. Vortex cDNA samples for 1 min, pulse-centrifuge, and incubate for 30 min at room temperature in the dark. 6. Repeat the previous step.

Identification of Transcription Factor Targets by Phenotypic Activation

29

7. Add 3.5 ml 4 M hydroxylamine to each cDNA sample to quench the reaction. Incubate for 15 min at room temperature in the dark. While the reaction is quenching, set up and label the Qiagen PCR purification columns and microcentrifuge tubes. 8. Purify labeled cDNA away from the dyes: Arrange the tubes in Cy3/Cy5 pairs – they will be combined in this step. 9. For all cDNA sample pairs: Add 70 ml water to the Cy3 cDNA sample. Add 500 ml buffer PB (Qiagen) to the Cy3 cDNA sample, mix by pipetting, combine with the Cy5 cDNA sample, and apply the mix to a Qiagen PCR purification column. 10. Centrifuge at 2,600 × g for 1 min and discard the flowthrough. 11. Apply 700 ml buffer PE (Qiagen), centrifuge at 2,600 × g for 1 min, and discard the flow-through. 12. Repeat the previous step. 13. Centrifuge at 16,000 × g for 1 min to dry the column. 14. Elute each cDNA sample from the column into a microcentrifuge tube twice with 30 ml elution buffer (Qiagen). The cDNA samples should be slightly purple in color. The efficiency of Cy3/Cy5 coupling to the cDNA samples can be determined at wavelengths 532 and 635 nm, respectively, with a spectrophotometer. 15. Dry the cDNA samples in a Speed-Vac (high-heat setting). Proceed to prehybridize the microarray slides. 3.6. Microarray Hybridization and Washing

1. Hybridization volumes, reagents, chambers, and conditions vary according to specific expression microarray platforms. The following protocol is used for expression microarrays manufactured by spotting 60-mer oligonucleotides (Open Biosystems) onto polylysine-treated glass microscope slides. 2. Microarray slide prehybridization: Thaw Quackenbush prehybridization buffer at 65°C and then incubate at 42°C. 3. UV-crosslink the microarray slides. 4. Place microarrays into slide mailers, fill with prehybridization buffer, and incubate at 42°C for a minimum of 45 min. 5. At the end of prehybridization, wash the microarrays 4–5 times with water by filling the slide mailer with water, rocking it, and pouring out the water. 6. Dry the microarrays by centrifuging them in slide racks in a benchtop centrifuge containing a swinging-bucket rotor and microplate carriers at 200 × g for 5 min. Store the microarrays

30

Chua

in a slide box at 42°C until ready to hybridize the Cy3/ Cy5-coupled cDNA samples. 7. Hybridizations: When the coupled cDNA samples are dry and the microarray slides are ready, resuspend each sample in 40 ml hybridization buffer. Add oligo spike-ins specific for control probes on the microarray if required. 8. Just prior to hybridization, heat the samples at 65°C for 3 min to denature, and then incubate at 42°C until ready for hybridization. 9. Place a microarray slide (array side up) at the bottom part of the chamber. 10. Pipette a Cy3/Cy5-coupled cDNA sample onto the microarray, taking care not to touch it with the pipette tip. Avoid the formation of bubbles since they will prevent uniform hybridization on the microarray. 11. Place one edge of the lifter cover slip on one side of the microarray slide with the lifter side down, and carefully lower the other edge onto the sample using a pair of curved forceps. Gently center the lifter cover slip over the microarray with forceps. 12. Fill the humidifying well at the end of the hybridization chamber with 4× SSC. 13. Seal the upper and lower portions of the hybridization chamber with screws/clips and incubate at 42°C overnight (or at least 6 h). 14. Record the label on the microarray and description of the samples in a hybridization sheet. 15. Repeat steps 9–14 for the next sample. 16. Microarray slide washes: After hybridization is complete, fill three 250-ml glass dishes with Wash 1, Wash 2, and Wash 3. 17. Open a hybridization chamber, and pick up a microarray slide with forceps. Submerge the microarray slide in Wash 1. The lifter cover slip should fall off almost immediately when the microarray slide is submerged. Be careful to avoid scratching the array. 18. Dip the microarray slide up and down 20 times in Wash 1. 19. Dip the microarray slide up and down 20 times in Wash 2. 20. Dip the microarray slide up and down 20 times in Wash 3. 21. Quickly dry the microarray slide (array side first) by immediately blowing off the wash solution with a dust gun or some other compressed gas, going from one side to the other in about 5 s. Store the microarray slide in a slide box. 22. Repeat steps 17–21 for the next microarray slide.

Identification of Transcription Factor Targets by Phenotypic Activation

1. Scan microarrays, quantitate images, and normalize data by Lowess smoothing with the laser scanner and integrated software according to the manufacturer’s instructions (7). Each pair of fluor-reversal experiments is combined by averaging the ratios of normalized median intensity of TF overexpression vs. the empty vector control for each spot on the microarray. 2. Examine the microarray data for specific upregulation of the TF mRNA in the corresponding TF overexpression experiment to confirm the induction of the TF gene in the presence of galactose and the correct orientation of the microarray samples (Fig. 2). 3. Calculate the Pearson correlation between each pair of fluorreversal experiments. A Pearson correlation of >0.3 between replicates is indicative that the differential gene expression in an experiment is more likely caused by TF overexpression rather than microarray noise (5). This correlation threshold may be different for other microarray platforms (Agilent, Affymetrix, Nimblegen, etc.).

log10 expression ratio

4. Input subsets of upregulated and downregulated genes separately at various thresholds of differential expression into FunSpec (http://funspec.med.utoronto.ca/) to identify significant functional enrichment of Gene Ontology categories (see Note 3). The segregation of upregulated and downregulated genes is required to characterize transcriptional activators and repressors, respectively. The significant functional

TF genes

3.7. Acquisition, Processing, and Analysis of Microarray Data

31

+2

0

–1.5

TF overexpression experiments

Fig. 2. Induction of transcription factor genes in 46 overexpression strains exhibiting phenotypic activation. The majority of transcription factor genes are highly induced in the corresponding transcription factor overexpression strains from microarray profiling (diagonal). The induction of transcription factor genes along the diagonal are highest and lowest in the upper left and lower right, respectively.

32

Chua

categories appearing in the list of differentially regulated genes should provide information on the TF’s putative function and targets since both are expected to be involved in a common biological process (Fig. 3). 5. Subject the microarray data of each TF overexpression experiment to motif searching using RankMotif (see Note 4). This probabilistic-inference algorithm is applied to the promoter regions of differentially expressed genes to identify prevalent 8-mer sequences that represent putative TF-binding sites (Fig. 3). Motif predictions can be validated using gel mobility shift assays involving the DNA-binding domain of the TF and two tandem copies of the predicted motif (3).

GCN4OE GCN4 MET6 HIS7 LYS1 TRP3 LEU1 ARG5 ASN1 HIS4 HIS5 CPA2 MET10 ECM40 ASN2 MET22 LYS9 ARG2 URA10 TRP2 LYS4 MET1 LYS14 CPA1 ARG1 ILV1 HIS1 ARG4 MET28 MET16 ARO9 ADE3 ECM17 ARG3 SER1 ARG8 HIS2 SER33 ILV6 GLN3 MET13 HOM2 ILV5 HIS3 ILV2 LYS12 TRP5 HOM3 ILV3 HIS6 ADE12

5'-RTGASTMA

Amino acid biosynthesis (P 50%

k n

k n S

MATa queryD::natMX4 xxxD::kanMX4 double mutant meiotic progeny

lethal

S lethal

lethal

Colony growth

Colony growth

Little or no Colony growth

Fig. 2. Genetic mapping using SGAM. In this example, the lethal phenotype associated with deletion of a query gene (n, green box) is suppressed by mutation of an unidentified gene (s, open box). SGA analysis is used to cross the query mutant harboring a suppressor mutation to the collection of ∼5,000 viable deletion mutants (k, purple box), all of which contain a wild-type copy of the unidentified gene (S, red box). As described in Subheading 3.3, double deletion mutants harboring the NatMX4-marked query deletion and a KanMX4-marked array deletion are selected following mating, meiotic recombination, and germination of haploid spore progeny. In the first example, the unidentified locus and KanMX4-marked deletion are located on different chromosomes and, therefore, segregate independently of one another. As a result, 50% of the resultant double deletion mutant meiotic progeny should contain the suppressor mutation locus and grow normally. In the second example, we expect that less than 50% of the NatMX4- and KanMX4-marked meiotic progeny will be viable because the array deletion mutant and the unidentified locus are linked and suppression of the growth defect is dependent on the distance separating the two genes and the frequency of meiotic recombination in the given chromosomal region. In the final example, the array-deletion strain and wild-type allele are tightly linked and, consequently, the number of viable double mutant progeny will be limited because of the reduced frequency of recombination between the suppressor allele and the KanMX4-marked strain. This results in the absence of colony growth. Adapted from (9) (see Color Plates).

pressor mutation(s) in an unidentified gene (S) is able to rescue the lethality associated with deletion of the query gene (9). The viable query strain harboring the suppressor (queryΔ::NatMX4s) is systematically crossed to each of the ∼5,000 haploid deletion array strains (xxxΔ::KanMX4) via SGA and double deletion meiotic progeny (queryΔ::NatMX4 xxxΔ::KanMX4) are selected. If the wild-type allele (S) and a particular array deletion mutant

48

Costanzo and Boone

(xxxΔ::KanMX4) are not linked, 50% of the double deletion meiotic progeny (queryΔ::NatMX4 xxxΔ::KanMX4) will contain the suppressor mutation (s) and germinate to form a colony. Conversely, recombination frequency is reduced when the deleted array gene (xxxΔ::KanMX4) and the wild-type allele (S) are tightly linked, resulting in limited recovery of viable double-deletion meiotic progeny. Hence, mapping the suppressor allele (s) involves identification of a colinear set of linked double mutants that fail to grow and form colonies. The chromosomal location of this linkage group defines the general location of the suppressor locus, which should occur roughly within the middle of the linkage group (see Note 14) (9). 3.4.1. Identifying Linkage Groups and Putative Suppressor Mutant Loci

1. Perform SGA screen using the “wild-type” control strain (Y8835) following the steps as described (see Subheading 3.3). 2. Visually inspect the experimental plates and compare to the wild-type control plates. Take note of the double mutant colonies that fail to grow, or appear smaller in size relative to wild-type controls (see Notes 13 and 14). 3. Record the potential interactions from the first round of screening. 4. Repeat the screen a second time. 5. Generate a set of putative interactions. While these data are expected to contain a substantial number of false positives, this unconfirmed SGA data should be sufficient for mapping purposes (see Notes 15 and 16). 6. Sort the set of putative interactions according to chromosome number and position (see step 7). 7. The putative suppressor mutation locus is determined via identification of a linkage group (i.e., failure to observe the formation of a potential collinear set of double mutant deletion strains; Fig. 3a; see Note 17). 8. The gene(s) that lie within the center of the linkage group should comprise the suppressor mutation. (Fig. 3b; see Note 18).

3.5. Confirming Identity of the Mutant Allele(s)

Once the candidate gene(s) is mapped, confirmatory genetic analysis is performed to prove that the mapped mutation is coincident with the suppressor mutation. 1. Sequence genes within the linkage group using standard techniques. Begin with genes situated closest to the center of the defined region, which should reveal the mutation (Fig. 3b). 2. Clone the mutant allele(s) from the suppressor strain using standard molecular biology techniques. 3. Identify the true suppressor allele(s) by complementation assays to prove that the mapped mutation is coincident with

SGAM: High-Resolution Genetic Mapping

A

49

WT screen

B

YPR003C YPR004C HAL1 ICL2 REC8 HAA1 SUT2 RPA135 YPR011C YPR013C YPR015C TIF6 DSS4 RLF2 CDC54 ATP20 AGC1 YPR022C EAF3 CCL1 ATH1 YPR027C

sdo1 D sup

ChrXVI

Fig. 3. SGAM identification of a suppressor of sdo1D slow growth. The SDO1 human ortholog has been linked to an autosomal recessive disorder known as the Shwachmann–Diamond syndrome. We previously isolated strains harboring a second-site mutation which suppressed the severe growth defect associated with deletion of SDO1 in yeast (11). (a) Example of a SGAM output plate consisting of 384 different sdo1D double mutants pinned in replicates of four. The KanMX4-marked array strains are arranged such that mutants corresponding to genes located adjacent to each other are arrayed at alternating positions of the same plate. A linkage group of inviable or slow-growing double mutants was identified following SGAM analysis (red squares) indicating the putative location of the sdo1D suppressor. (b) Genes located within the chromosome 15 linkage group are shown. Red bars indicate slow growth or death of the particular deletion when combined with the sdo1D mutant. Black bars indicate essential genes not represented in the arrayed collection of KanMX4-marked strains. Sequencing of alleles within this linkage group identified a mutation in the TIF6 gene (located near the center of the linkage group). Indeed, additional experiments confirmed that a dominant TIF6 gain-of-function mutation rescues the slow growth phenotype associated with a sdo1D strain. Adapted from (11) (see Color Plates).

the suppressor mutation. Introduce cloned alleles and vector controls into the appropriate deletion mutant strain and identify the cloned allele(s) that restores wild-type growth to that of the original slow-growing deletion mutant. 3.6. SGAM Applications

In principle, SGAM can be applied to uncover modifiers of any allele that leads to a phenotype distinct from the wild-type strain (9). These phenotypes may include conditional alleles of essential genes, extragenic suppressors of null mutants, expression of chromophores (or other reporter genes), as well as cytotoxic drug resistance. In addition to mapping of recessive alleles, SGAM

50

Costanzo and Boone

is particularly useful for rapid mapping of dominant mutations that are challenging to clone using standard techniques (11). Moreover, combinations of alleles responsible for complex, multigenic quantitative traits could be mapped in a single SGAM experiment. SGAM has been used successfully to map locations of several extragenic suppressors of yeast deletion alleles (9– 11). Notably, SGAM was recently used to identify a dominant suppressor of a gene associated with the Shwachman–Diamond syndrome. This study provided insight into the molecular basis of this disease and emphasized the importance of genetic analysis in model organisms (11).

4. Notes 1. When making up the amino acid supplement mixture, the solid ingredients should be combined and then mixed thoroughly by turning end-over-end for at least 15 min. The resultant mixture can be stored in tinted glass bottles at room temperature. 2. Ammonium sulfate impedes the function of G418 and clonNAT. Hence, synthetic medium containing these antibiotics is made with monosodium glutamic acid as a nitrogen source (13). 3. This medium does not contain any antibiotics such as G418 and clonNAT and therefore ammonium sulfate is used as the nitrogen source. 4. The Singer RoTor DHA benchtop robot uses disposable replicators (RePads, Singer Instruments, UK) and PlusPlates (Singer Instruments, UK) that have a larger surface area but the same external footprint dimensions as OmniTray (Nunc). The BioMatrix robot can be used in conjunction with OmniTrays for the replica pinning steps involved in SGA analysis. We use 100-mm Petri dishes for the construction of SGA query strains and tetrad analysis, and 60-mm Petri dishes for random spore analysis. We found that approximately 35 and 50 mL of media in OmniTrays and PlusPlates, respectively, yield optimal results. For random spore analysis, approximately 10 mL of media in a 60-mm dish is optimal. 5. The 1.58-mm diameter, flat-tip pins (FP6) can be used as an alternative to the chamfered-tip pins. However, they transfer more cells than the chamfered-tip pins and may not be suitable for producing high-density arrays (1,536 spots/array). 6. Empty tip boxes can be used as a substitute to the reservoirs for bleach, water, and ethanol.

SGAM: High-Resolution Genetic Mapping

51

7. We have also constructed ordered arrays comprising temperature-sensitive alleles of essential genes. 8. To ensure that the pins are cleaned properly and to avoid contamination in the wash procedure, the volume of wash liquids in the cleaning reservoirs is calculated to cover the pins sequentially in small increments. For example, only the tips of the pins should be submerged in water in the first step. As the pins are transferred to subsequent cleaning reservoirs and the final ethanol step, the lower halves of the pins should be covered. 9. Deletion mutant arrays are frozen in 96-well microtiter plates. It is feasible to conduct SGA and SGAM at this density (96 mutants/plate). However, screen throughput and accuracy are increased by construction of high-density deletion arrays. Array densities include, 384, 768, or 1,536 colonies/plate. Methods for assembling higher density arrays have been described in detail (14). 10. Pinning the query strain in a 1,536-format on an agar plate is advantageous, as cells are evenly transferred to subsequent mating steps. One query plate should contain a sufficient amount of cells for mating with eight plates of the DMA. 11. The DMA can be reused for three to four rounds of mating. 12. It is important to incubate the sporulation plates at approximately 22–24 °C for efficient sporulation. Following sporulation, these plates can be stored at 4 °C for up to 4 months without significant loss of spore viability, to provide a source of spores for subsequent random spore analysis and tetrad analysis. 13. In addition to visual inspection of the double mutants, we have developed a computer-based scoring system, which generates an estimate of relative growth rates from the area of individual colonies as measured from digital images of the double-mutant plates. Statistical significance can be determined for each strain by comparing the measurements between the mutants and wild-type controls. 14. SGAM provides several advantages as a high-resolution genetic mapping approach. First, although the mutation is mapped by linkage to nonessential genes, it may lie in either nonessential or essential genes. Second, both dominant and recessive mutations can be mapped. Third, as the size of the linked group will be inversely proportional to the frequency of meiosis in the diploid colonies, the assay could be finetuned by modulating sporulation efficiency. For instance, hampering sporulation by manipulating temperature and/ or incubation times should extend the linked region.

52

Costanzo and Boone

In addition, multigenic traits in which several alleles are required to confer the mutant phenotype can also be mapped (9). 15. Although SGAM is applied for mapping mutations, it is possible to identify synthetic lethal or sick interactions between the query gene deletion mutant and the array of 5,000 viable deletion mutants from the same SGAM screen. However, tetrad analysis is required in order to ensure that observed synthetic lethal/sick interactions are specific to combinations of array and query mutations and not to the unknown suppressor mutation(s). 16. Linkage groups can usually be identified from raw SGA data, and therefore random spore and tetrad analysis is not required to map suppressor loci. However, random spore and tetrad analysis can be used to determine the frequency of recombination between the putative suppressor allele and KanMX4-marked array gene deletion strains. Random spore and tetrad analyses have been previously described (14). 17. The number of linkage groups identified by SGAM is dependent on the nature of the genetic trait. If the suppression phenotype is attributable to a single locus, then we expect to find two linkage groups: one corresponding to the NatMX4-marked query mutation and a second corresponding to the suppressor allele. We expect to identify more than two linkage groups if the suppression phenotype is a multigenic trait. 18. We have developed an algorithm for detecting significant linkage groups in SGAM (9).

Acknowledgments We thank Anastasia Baryshnikova, Renee Brost, Julie Guzzo, and Corey Nislow for help with figures and comments on the manuscript. This work was supported by grants to CB from the Canadian Institutes for Health Research and Genome Canada through the Ontario Genomics Institute.

SGAM: High-Resolution Genetic Mapping

53

References 1. Boone, C., Bussey, H., et al. (2007). Exploring genetic interactions and networks with yeast. Nature Reviews Genetics 8, 437–449. 2. Giaever, G., Chu, A. M., et al. (2002). Functional profiling of the Saccharomyces cerevisiae genome. Nature 418, 387–391. 3. Winzeler, E., Shoemaker, D. D., et al. (1999). Functional Characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science 285, 901–906. 4. Kaelin, W. G. (2005). The concept of synthetic lethality in the context of anticancer therapy. Nature Reviews Cancer 5, 689–698. 5. Hartman, J. L., Garvik, B., et al. (2001). Principles for the buffering of genetic variation. Science 291, 1001–1004. 6. Hartwell, L. (2004). Genetics. Robust interactions. Science 303, 774–775. 7. Tong, A., Lesage, G., et al. (2004). Global mapping of the yeast genetic interaction network. Science 303, 808–813. 8. Tong, A. H. Y., Evangelista, M., et al. (2001). Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294, 2364–2368.

9. Jorgensen, P., Nelson, B., et al. (2002). Highresolution genetic mapping with ordered arrays of Saccharomyces cerevisiae deletion mutants. Genetics 162, 1091–1099. 10. Chang, M., Bellaoui, M., et al. (2005). RMI1/NCE4, a suppressor of genome instability, encodes a member of the RecQ helicase TopoIII complex. EMBO J. 24, 2024–2033. 11. Menne, T. F., Goyenechea, B., et al. (2007). The Shwachman-Bodian-Diamond syndrome protein mediates translation activation of ribosomes in yeast. Nature Genetics 39, 486–495 12. Goldstein, A. L., McCusker, J. H. (1999). Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15, 1541–1553. 13. Cheng, T. H., Chang, C. R., et al. (2000). Controlling gene expression in yeast by inducible site-specific recombination. Nucleic Acids Research 28, E108. 14. Tong, A., Boone, C. (2006). Synthetic genetic array analysis in Saccharomyces cerevisiae. In Xiao W (ed.) Yeast Protocols, 2nd edn. Humana, Totowa, NJ, pp. 171–191.

Chapter 4 Reporter-Based Synthetic Genetic Array Analysis: A Functional Genomics Approach for Investigating the Cell Cycle in Saccharomyces cerevisiae Holly E. Sassi, Nazareth Bastajian, Pinay Kainth, and Brenda J. Andrews Summary Temporal control of gene expression is a widespread feature of cell cycles, with clear transcriptional programs in bacteria, yeast, and metazoans. In budding yeast, approximately 1,000 genes are transcribed during a specific interval of the cell cycle. Although a number of factors that contribute to this periodic pattern of gene expression have been studied in Saccharomyces cerevisiae, pathways of cell cycle-regulated transcription remain largely undefined. To identify regulators of genes exhibiting cell cycle periodicity, we have developed a functional genomics approach termed reporter-based synthetic genetic array (R-SGA) analysis. Based on synthetic genetic array (SGA) analysis, R-SGA allows rapid and easily automated incorporation of a cell cycle reporter gene into the array of viable haploid yeast gene-deletion mutants. Scoring of reporter activity in mutant strains compared to wild type identifies candidate regulators of the cell cycle gene of interest. In contrast to microarrays, which generally provide information about the expression of all genes under a particular condition (for example, a single gene deletion), R-SGA analysis facilitates the study of the expression of a single gene in all deletion mutants. Our system can be adapted to examine the expression of any gene not only in the context of haploid deletion mutants but also using other array-based strain collections available to the yeast community. Key words: Yeast, Synthetic genetic array, Reporter, High-throughput plasmid transfer, Cell cycle, Transcription, lacZ, HIS3, 3-Aminotriazole

1. Introduction Cells must follow a specific genetic program to ensure high-fidelity chromosome duplication and segregation. The cell division cycle of all eukaryotes consists of four main phases: DNA replication I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_4, © Humana Press, a part of Springer Science + Business Media, LLC 2009

55

56

Sassi et al.

(S phase) and mitosis (M phase), separated by gap phases (G1 and G2). In Saccharomyces cerevisiae, the cell cycle is characterized by dramatic periodic alternations of gene expression, with approximately 15% of all genes fluctuating in successive transcriptional waves (1–3). Key cell cycle regulators, including the mitotic cyclins, transcription factors, and DNA replication complex components, exhibit periodic patterns of expression along with many genes that carry out cell cycle-specific functions. The set of genes causing fitness defects when overexpressed is enriched for cell cycle-regulated genes (4), highlighting the importance of their restricted expression. Transcriptional profiling (1–3), genome-wide chromatin immunoprecipitation (ChIP-chip) (5–8), as well as bioinformatic (9, 10) and mechanistic studies have collectively identified hundreds of potential targets and regulators of cell cycle-regulated transcription factors. However, the global picture of cell cycle circuitry is far from complete. For example, at the G1/S transition alone, only 23% of the promoters adjacent to G1-phase genes and 12% of the promoters adjacent to S-phase genes appear to be bound by SBF or MBF (5), the known transcriptional regulators of this transition. We have developed reporter-based synthetic genetic array (R-SGA) analysis, a novel method that facilitates the systematic identification of genes that regulate expression from cell cycleregulated promoters (CCprs) in vivo. In R-SGA, a query yeast strain bearing a CCpr fused to a reporter gene is mated to the haploid deletion mutant array (11–13). The query strain contains selectable markers that allow the use of the SGA method (12, 14, 15) to obtain haploid cells that contain both the reporter gene and deletion mutation through a series of robotic pinning steps. This plasmid transfer protocol effectively replaces thousands of manual yeast transformations. The resulting strains are assayed for reporter activity; enhanced reporter activity in a particular deletion mutant indicates that the deleted gene is a putative repressor of the cell cycle-regulated gene, while a reduction in reporter activity indicates a candidate activator. A control promoter (Controlpr), which is not cell cycle regulated, is used to identify mutants with nonspecific transcriptional defects. We describe the system using plasmid-based URA3-marked Escherichia coli lacZ and S. cerevisiae HIS3 reporters crossed to the array of viable haploid deletion mutants; however, with minor modifications our approach could be adapted to incorporate any reporter in either a plasmid or integrated context, any compatible selectable marker, and any collection of arrayed yeast strains.

Reporter-Based Synthetic Genetic Array Analysis

57

2. Materials 2.1. Reporters

Starting vectors for reporter gene construction (see Subheading 3.1): 1. pΔSS (16). 2. pRS423, pRS426 (17).

2.2. R-SGA Procedure

Yeast gene-deletion mutant collection maintenance and all R-SGA selection steps are carried out in OmniTray plates (Nunc, Catalog #242811) using manual or robotic pinning tools as described in refs.12, 14, and 15. Standard yeast manipulations are performed in 92 × 16 mm petri dishes (Sarstedt). 1. The yeast gene-deletion mutant collection is derived from strain BY4741 (MATa ura3D0 leu2D0 his3D1 met15D0) (18). The collection consists of all S. cerevisiae viable haploid deletion mutants in which a given open reading frame (ORF), XXX, has been replaced by a kanamycin resistance marker (kanr); xxxD::kanr(11, 13). The collection is available from Invitrogen (http://www.invitrogen.com), American Type Culture Collection (http://www.atcc.org), EUROSCARF (http://www.uni-frankfurt.de/fb15/micro/euroscarf/ index.html), and Open Biosystems (http://www.openbiosystems.com/GeneExpression/Yeast). 2. Y7039 (MATα can1Δ::STE2pr-LEU2 lyp1Δ ura3Δ0 leu2Δ0 his3Δ1 met15Δ0) (15) derived from BY4741 (18). 3. Standard yeast transformation reagents. 4. Glucose (Dextrose, Sigma-Aldrich Co.): Prepare a 50% (w/v) solution in distilled water, autoclave, and store at room temperature. 5. 10× amino acid supplements for synthetic media (complete): Add 300 mg isoleucine, 1.5 g valine, 400 mg adenine hemisulfate salt, 200 mg arginine hydrochloride, 200 mg histidine, 1 g leucine, 200 mg uracil, 300 mg lysine monohydrochloride, 200 mg methionine, 500 mg phenylalanine, 2 g threonine, 400 mg tryptophan, and 300 mg tyrosine (Sigma-Aldrich Co.) to a beaker. Bring the volume up to 1 L with distilled water, mix, autoclave, and store at 4°C. 10× amino acid drop-out (DO) solution is a combination of the above ingredients minus the appropriate supplement. For the R-SGA protocol described, 10× −Ura and 10× −Ura/ Leu/Arg/Lys are required. 6. Amino acid supplement for enriched sporulation media: Dissolve 0.714 g uracil, 0.714 g histidine, and 3.571 g

58

Sassi et al.

leucine in 500 mL of distilled water and filter-sterilize. Store the solution at 4°C. 7. G418 (G-418 sulfate/geneticin, Gibco, Invitrogen Corp.): Dissolve in distilled water at 200 mg/mL, filter-sterilize, and store stock solutions in aliquots at 4°C. Store the powder at room temperature. 8. Canavanine (L-canavanine sulfate salt, Sigma-Aldrich Co.): Dissolve in distilled water at 50 mg/mL, filter-sterilize, and store stock solutions in aliquots at 4°C. Store the powder at 4°C. 9. Thialysine (S-(2-aminoethyl)-L-cysteine hydrochloride/L4-thialysine hydrochloride, Sigma-Aldrich Co.): Dissolve in distilled water at 50 mg/mL, filter-sterilize, and store stock solutions in aliquots at 4°C. Store the powder at 4°C. 10. SD −Ura (synthetic minimal glucose medium supplemented with 10× −Ura DO solution): Add 6.7 g of yeast nitrogen base without amino acids (BD Difco Laboratories) and 860 mL of distilled water to a 2 L flask, mix, and autoclave. Add 40 mL of 50% glucose and 100 mL of 10× −Ura amino acid DO solution and mix. For plates, add 20 g agar (BD Difco Laboratories) prior to autoclaving. 11. YEPD (yeast extract–peptone–dextrose) plates: Add 120 mg adenine, 10 g yeast extract, 20 g peptone (BD Difco Laboratories), and 20 g agar to a 2 L flask, bring the volume up to 960 mL with distilled water, mix, and autoclave. Add 40 mL of 50% glucose and mix thoroughly. 12. YEPD + G418 plates (see Note 1): Cool the YEPD medium to ∼65°C, add 1 mL of G418 stock solution/L (final concentration 200 mg/L), and mix. 13. (SD/MSG)-Ura + G418 plates (see Note 2): Add 1.7 g yeast nitrogen base without amino acids and ammonium sulfate (BD Difco Laboratories), 1 g monosodium glutamic acid (MSG), Sigma-Aldrich Co., and 20 g agar to a 2 L flask, bring the volume up to 860 mL with distilled water, mix, and autoclave. Add 40 mL of 50% glucose and 100 mL of 10× −Ura amino acid DO solution. Cool to ∼65°C, add 1 mL of G418 stock solution (final concentration 200 mg/L), and mix thoroughly. 14. Enriched sporulation + G418 plates (see Note 1): Add 10 g potassium acetate (BioShop Canada Inc.), 1 g yeast extract, 0.5 g glucose, and 20 g agar to a 2 L flask. Bring the volume up to 990 mL with distilled water, mix, and autoclave. Add 10 mL of amino acid supplement solution and cool to ∼65°C. Add 250 μl of G418 stock solution (final concentration 50 mg/L) and mix thoroughly. 15. SD −Ura/Leu/Arg/Lys + canavanine/thialysine plates: Add 6.7 g yeast nitrogen base without amino acids and

Reporter-Based Synthetic Genetic Array Analysis

59

20 g agar to a 2 L flask. Add 860 mL of distilled water, mix, and autoclave. Add 40 mL of 50% glucose and 100 mL of 10× −Ura/Leu/Arg/Lys amino acid DO solution. Cool to ~65°C, add 1 mL of canavanine (can) stock solution (final concentration 50 mg/L) and 1 mL of thialysine (thia) stock solution (final concentration 50 mg/L), and mix thoroughly. 16. (SD/MSG)-Ura/Leu/Arg/Lys + can/thia/G418 plates (see Note 2). Add 1.7 g yeast nitrogen base without amino acids and ammonium sulfate, 1 g MSG, and 20 g agar to a 2 L flask. Bring the volume up to 860 mL with distilled water, mix, and autoclave. Add 40 mL of 50% glucose and 100 mL of 10× −Ura/Leu/Arg/Lys amino acid DO solution. Cool to ~65°C, add 1 mL of can stock solution (final concentration 50 mg/L), 1 mL of thia stock solution (final concentration 50 mg/L), and 1 mL of G418 stock solution (final concentration 200 mg/L) and mix thoroughly. 2.3. Identification of CCpr-lacZ Reporter Regulators 2.3.1. Array-Format b-Galactosidase Activity Assay

1. SD −Ura plates. 2. Agarose (BioShop Canada, Inc.). 3. DMF (N,N-dimethylformamide, Sigma-Aldrich Co.). DMF is toxic and harmful by inhalation or on contact with skin. Work in a fume hood and wear protective clothing. 4. SDS (sodium dodecyl sulfate, Bio-Rad Laboratories, Inc.): Add 100 g to 900 mL of distilled water and heat to 68°C to dissolve. Adjust the pH to 7.2 by adding a few drops of concentrated HCl. Adjust the volume to 1 L with distilled water (10% stock solution) and store in aliquots. A mask should be worn when weighing SDS. 5. X-gal (5-bromo-4-chloro-3-indolyl-β-galactopyranoside, Sigma-Aldrich Co.): Make a 20 mg/mL stock solution by dissolving in DMF. Store in the dark at −20°C. 6. KPO4 (0.5 M potassium phosphate buffer, pH 7.0): Mix appropriate volumes of 0.5 M K2HPO4 (∼61.5 mL) and 0.5 M KH2PO4 (∼38.5 mL) such that the final pH of the solution is 7.0. 7. 50 mL conical tubes (BD Biosciences). 8. Hot plate with magnetic stirrer.

2.3.2. Confirmation of Candidate CCpr-lacZ Regulators

1. YEPD plates. 2. YEPD + G418 plates. 3. Standard yeast transformation reagents. 4. SD −Ura plates. 5. X-gal top agarose solution (see Subheadings 2.3.1 and 3.3.1).

60

Sassi et al.

2.4. Identification of CCpr-HIS3 Reporter Regulators 2.4.1. Array-Format HIS3 Activity Assay

1. 10× −His amino acid DO solution. 2. 3-AT (3-amino-1,2,4-triazole/3-aminotriazole, Sigma-Aldrich Co.): Make a 2.5 M stock in distilled water, filter-sterilize, and store at 4°C. 3-AT is possibly carcinogenic; work in a fume hood and wear protective clothing. 3. SD −His + 3-AT plates: Add 6.7 g yeast nitrogen base without amino acids and 20 g agar to a 2 L flask, bring the volume up to 860 mL with distilled water, mix, and autoclave. Add 40 mL of 50% glucose and 100 mL of 10× −His amino acid DO solution. Cool to ~65°C, add the appropriate volume of 3-AT stock (for 10 or 75 mM final concentration), and mix thoroughly.

2.4.2. Confirmation of Candidate CCpr-HIS3 Regulators

1. YEPD + G418 plates. 2. Standard yeast transformation reagents. 3. 96-well round bottom assay plates (Costar #3795). 4. 96-well microplate lids (Costar #3930). 5. SD −Ura liquid medium. 6. SD −Ura plates. 7. SD −His + 10 mM 3-AT plates. 8. SD −His + 75 mM 3-AT plates.

3. Methods 3.1. Reporters

To generate the parent HIS3 reporter construct, the S. cerevisiae HIS3 gene was amplified from pRS423 (17) using primers HIS3 Left (5¢ CGAGGATCCGGCAAAGATGACAGAGCAGAA 3¢) and HIS3 Right (5¢ GCAGAATTCACCGCATAGATCCGTCGAGT 3¢), which contain engineered BamHI and EcoRI sites, respectively. The PCR fragment was digested with BamHI and EcoRI and cloned into similarly digested pRS426 (17). The lacZ reporter constructs used in our R-SGA studies are primarily derivatives of pΔSS (16). Both pΔSS and pHIS3 contain a URA3 marker, the yeast 2-µm origin of replication (see Note 3), and an ampicillin resistance marker (ampr). Any desired CCpr or specific promoter element can be amplified from genomic/other DNA using PCR primers containing engineered restriction sites for subcloning into either the lacZ or HIS3 parent reporter plasmid (see Notes 4 and 5). Control reporter plasmids used in confirmation assays to identify noncell-cycle-specific transcriptional effects contain either the S. cerevisiae RPL39 or ACT1 promoter (see Note 6).

Reporter-Based Synthetic Genetic Array Analysis

3.2. R-SGA Procedure

61

An overview of the R-SGA method is shown in Fig. 1a. All R-SGA selection steps are performed in 768-density array format (see Note 7). For protocols to generate the 768-array format yeast gene-deletion set from 96-well plates, array maintenance, sterilization of robotic and manual pinning tools, and manual pinning protocols, see refs. 12, 14, and 15. The following procedure facilitates high-throughput transfer of the CCpr-reporter plasmid from strain Y7039 to each xxxΔ::kanr strain in the yeast gene-deletion set (see Note 8). Adaptations of the method are described in Subheading 4 (see Note 9). 1. To generate the R-SGA query strain, introduce the CCpr-reporter plasmid into strain Y7039 by standard yeast transformation, plate on SD −Ura, and incubate for 2–3 days at 30°C. 2. Patch the transformants onto SD −Ura and incubate for 2 days at 30°C. 3. Inoculate the query strain into 10 mL of liquid SD −Ura. Grow to saturation at 30°C overnight. 4. Pellet the culture (2,100 × g, 5 min), resuspend in 5 mL of fresh SD −Ura, and plate onto two YEPD plates (OmniTray format). Grow for 2 days until dense lawns of yeast have formed (see Note 10). 5. In parallel, propagate the yeast gene-deletion set in 768array format at 30°C on YEPD + G418 plates. 6. Pin the query strain containing the CCpr-reporter plasmid from the lawns onto YEPD plates. Pin the yeast gene-deletion set directly on top of the query cells to allow the cells to mate. Incubate for 1 day at 30°C. 7. Pin the resulting MATa/α diploids onto (SD/MSG) −Ura + G418 to select for the plasmid and the xxxΔ::kanr gene, respectively. Incubate for 2 days at 30°C. 8. Repeat the above diploid selection pinning step and incubate the plates at 30°C for 1 day. 9. Pin the diploid cells onto enriched sporulation medium + G418. Incubate the sporulation plates for 5–9 days in the dark at 22°C. 10. Pin the spores onto SD −Ura/Leu/Arg/Lys + can/thia plates in duplicate (see Note 11) to select for MATa meiotic progeny carrying the CCpr-reporter plasmid. Incubate the haploid selection plates at 30°C for 2 days. 11. Pin onto (SD/MSG) −Ura/Leu/Arg/Lys + can/thia/G418 to select for MATa xxxΔ::kanr cells carrying the CCpr-reporter plasmid. Incubate the final selection plates at 30°C for 2 days. 12. Repeat the final selection pinning on (SD/MSG) −Ura/ Leu/Arg/Lys + can/thia/G418. Incubate the plates for 2 days at 30°C.

62

Sassi et al.

a xxxΔ

X

XXX MATα

Mating

YEPD

MATa

Selection for diploids containing CCpr-reporter plasmid

XXX xxxΔ a/α

XXX

G)-Ura (SD/MS + G418

XXX

Sporulation xxxΔ

xxxΔ

xxxΔ

tion

Sporula

Selection for MATa haploids containing CCpr-reporter plasmid

XXX

xxxΔ

Selection for MATa xxxΔ haploids containing CCpr-reporter plasmid

b

c

SD-Ura

rg / Lys

/Leu /A SD-Ura ia th / n a c +

rg /Leu /A G)-Ura S /M D (S 8 ia /G41 + can/th

SD-His +

/ Lys

AT

10 mM 3-

X-gal top agarose solution

SD-Ura

SD-His +

AT

75 mM 3-

Fig. 1. Overview of the R-SGA methodology. (a) R-SGA selection steps. The MATα query strain containing the URA3marked CCpr-reporter plasmid is crossed to the ordered array of MATa viable haploid gene-deletion mutants on YEPD. Each mutant carries a gene deletion linked to a kanr marker that confers resistance to G418. The resulting MATa/α diploids are pinned onto medium lacking uracil and supplemented with G418 to select the plasmid and gene deletion linked to the kanr marker, respectively. Diploids are transferred to enriched sporulation medium to induce sporulation and the formation of haploid meiotic spore progeny. Spores are subsequently transferred to medium lacking uracil, leucine, arginine, and lysine and supplemented with canavanine and thialysine. The omission of uracil maintains plasmid selection. The omission of leucine allows selective growth of MATa meiotic progeny, since only these cells express the STE2prLEU2 reporter. can1Δ and lyp1Δ markers and canavanine and thialysine selections are described in Tong and Boone (15). Strains are subsequently pinned onto medium lacking uracil, leucine, arginine, and lysine and supplemented with canavanine, thialysine and G418 to select for MATa gene-deletion mutant cells containing the CCpr-reporter plasmid. XXX: wild-type allele, xxxΔ: gene-deletion allele, black oval: URA3-marked plasmid, black box: cell cycle-regulated promoter,

Reporter-Based Synthetic Genetic Array Analysis

63

See Subheadings 2.3 and 3.3 for CCpr-lacZ constructs or Subheadings 2.4 and 3.4 for CCpr-HIS3 constructs. Fig. 1b, c depicts downstream reporter activity assays for lacZ and HIS3 constructs, respectively. 3.3. Identification of CCpr-lacZ Reporter Regulators

1. Pin the array of deletion mutants containing the CCpr-lacZ plasmids a final time onto SD −Ura. Incubate for 1–2 days at 30°C.

3.3.1. Array-Format b-Galactosidase Activity Assay

2. Prepare top agarose to cover the genome-wide array (this recipe is enough for 18 OmniTray plates): a. Add 1.8 g of agarose (final concentration 0.5%) to a 1 L flask. b. Add 334.8 mL of 0.5 M KPO4. c. Mix, boil, and cool to 60°C on a hot plate with stirring. d. Add 14.4 mL of DMF and 3.6 mL of 10% SDS and mix thoroughly. 3. For one OmniTray plate, dispense 19.6 mL of the top agarose solution into a 50 mL conical tube and add 400 μl of 20 mg/mL X-gal in DMF. Invert to mix. 4. Pour the X-gal top agarose over the plate on a flat surface, and immediately tip the plate to distribute the solution evenly over the surface of the agar (see Note 12). 5. Repeat steps 3 and 4 for the remaining plates. Incubate the plates at 37°C for 1 h (see Note 13). 6. Score deletion mutants containing the CCpr-lacZ plasmid for β-galactosidase activity. Deletion of putative activators will result in white colonies with no β-galactosidase activity, while deletion of repressors will result in dark blue colonies with enhanced reporter activity (see Note 14). An example of results produced is shown in Fig. 2a. Liquid β-galactosidase assays may be used if quantitative results are desired (Fig. 2b).

Fig. 1. (continued) grey box: reporter. Adapted from Tong and Boone (15). (b) Detection of lacZ reporter activity. Yeast gene-deletion mutants containing a CCpr-lacZ plasmid are pinned onto medium lacking uracil. X-gal top agarose solution is poured over the plates and β-galactosidase activity (i.e., the extent of blue color) is scored. Black arrows indicate putative regulators. Black circle: yeast strain exhibiting increased β-galactosidase activity in which a candidate repressor is deleted, white circle: yeast strain exhibiting decreased β-galactosidase activity in which a candidate activator is deleted. (c) Detection of HIS3 reporter activity. Yeast gene-deletion mutants containing a CCpr-HIS3 reporter are pinned onto medium lacking histidine and supplemented with low or high concentrations of 3-AT. Black arrows on the SD −His +10 mM 3-AT plate indicate strains that are unable to grow, which contain deletions in candidate activators. Black arrows on the SD −His +75 mM 3-AT plate indicate strains that are able to grow in the presence of high concentrations of 3-AT, which contain deletions in candidate repressors.

64

Sassi et al.

b a

Fig. 2. Identifying regulators of a G1-phase cell cycle promoter element using a lacZ reporter. (a) Identification of MCB-lacZ regulators in array format. A reporter construct containing four copies of the MluI cell cycle box (MCB) sequence element fused to lacZ (19) was crossed to the yeast gene-deletion array, and haploid MATa xxxΔ::kanr strains containing the MCBlacZ plasmid were isolated using the R-SGA selection steps. The strains were pinned onto SD −Ura plates, grown for 2 days, and subsequently covered with top agarose containing X-gal to identify regulators of MCB-lacZ expression. As expected, deletion of SWI6 and MBP1, the components of the MBF transcription factor that binds to MCB elements and activates transcription of G1-specific genes, resulted in a failure to produce β-galactosidase. A 384-array format is shown. (b) Liquid β-galactosidase assays. RPL39-lacZ (see Note 6) or MCB-lacZ reporter plasmids were introduced into swi6Δ, mbp1Δ, and wild-type strains, and quantitative liquid β-galactosidase assays were performed to confirm screen results. MCB-lacZ expression was reduced in swi6Δ and mbp1Δ strains compared to wild type, and the effect was specific to MCB elements. Values for β-galactosidase activity were calculated relative to wild-type levels for each plasmid (see Color Plates).

3.3.2. Confirmation of Candidate CCpr-lacZ Regulators

1. From the original master deletion set plate, streak the xxxΔ::kanr strains corresponding to the mutants identified in the primary screen on YEPD + G418 and streak BY4741 (wild type) on YEPD. Incubate for 2 days at 30°C. 2. Introduce the CCpr-lacZ plasmid into the deletion strains and into BY4741 by standard yeast transformation, and plate onto SD −Ura. Incubate for 2–3 days at 30°C. 3. Pick four transformants per gene-deletion strain and patch them onto SD −Ura plates in a 96-density array format. Incubate for 2 days at 30°C. 4. Pin the 96-density array onto a fresh SD −Ura plate. Incubate for 2 days at 30°C. Retain this plate as the master plate of candidate regulator gene-deletion strains identified in the screen. 5. In parallel, make a lawn of strain BY4741 containing the CCpr-lacZ plasmid on SD −Ura. Incubate the lawn for 2 days at 30°C. 6. Pin from the lawn onto a fresh SD −Ura plate in 96-array format. Incubate for 2 days at 30°C. 7. Pin the 96-density array of candidate regulator deletion mutants containing the CCpr-lacZ plasmid onto a fresh SD −Ura plate in 192-array format. 8. Pin the BY4741 colonies containing the CCpr-lacZ plasmid onto the same SD −Ura plate in the offset position in 192-array

Reporter-Based Synthetic Genetic Array Analysis

65

format. This will generate a 384-density array in which all candidate regulator gene-deletion strains are adjacent to wildtype colonies (Fig. 3). Incubate for 1–2 days at 30°C.

WT

Introduce CCpr-lacZ plasmid into candidate xxxΔ and wild-type strains

xxxΔ

Make WT containing CCpr-lacZ lawn

Patch 4 transformants in 96-array format

Pin in 96-array format

WT containing CCpr-lacZ

Deletion mutants containing CCpr-lacZ

Pin onto a single plate in 384-array format

Fig. 3. Building a 384-format array to confirm regulators identified in genome-wide lacZ R-SGA screens. The CCpr-lacZ plasmid is introduced into the wild-type strain and into all yeast gene-deletion strains identified as candidate regulators in the primary R-SGA screen. A lawn of the wild-type strain containing the CCpr-lacZ plasmid is made and a 96-head pinning tool is used to array the strain. In parallel, four independent transformants of each xxxΔ::kanr strain containing the CCprlacZ plasmid are patched in a 96-array format (boxes). These strains are subsequently pinned onto a fresh plate in duplicate offset positions resulting in a 192-format array of 24 distinct plasmid-containing deletion mutants, each represented eight times (box on 384-array plate). A 96-head pinning tool is used to transfer the wild-type strains containing the plasmid onto the same plate in different offset positions, resulting in 192 wild-type colonies. Each deletion strain (blue) is adjacent to a wild-type colony (light grey) on the final 384-format array plate for direct comparison of CCpr-lacZ activity. All plates are SD −Ura for plasmid selection. This procedure is carried out in parallel with the Controlpr-lacZ construct to determine the specificity of candidate regulators for the CCpr-lacZ reporter. Blue rectangle: lacZ, WT: wild type (see Color Plates).

66

Sassi et al.

9. Perform the X-gal overlay assay as described (see Subheadings 2.3.1 and 3.3.1). 10. Repeat steps 2–9 using the Controlpr-lacZ construct. 11. Score the xxxD::kanr and BY4741 strains containing the CCpr-lacZ construct by comparing mutants next to wild type to identify high-confidence candidate regulators (see Note 15). 12. Score the xxxΔ::kanr and BY4741 strains containing the Controlpr-lacZ construct in the same way (see Note 16). 13. Identify the candidate regulators that are specific to the CCpr-lacZ construct (see Notes 17 and 18). 3.4. Identification of CCpr-HIS3 Reporter Regulators 3.4.1. Array-Format HIS3 Activity Assay

1. Pin the array of deletion mutants containing the CCpr-HIS3 plasmids a final time onto SD −His + 10 mM 3-AT and SD −His + 75 mM 3-AT plates (see Note 19). 2. Incubate the plates for 2 days at 30°C. 3. Score the SD −His + 10 mM 3-AT plates for colonies that are slow growing or dead compared to those growing on the (SD/MSG) −Ura/Leu/Arg/Lys + can/thia/G418 final selection plates. These strains contain deletions in candidate activators (see Note 20). 4. Identify colonies on the SD −His + 75 mM 3-AT plates that are able to grow. These strains contain deletions in candidate repressors. Results from a genome-wide R-SGA screen are depicted in Fig. 4a, b.

3.4.2. Confirmation of Candidate CCpr-HIS3 Regulators

1. From the original master deletion set plate, streak the xxxΔ::kanr strains corresponding to the mutants identified in the primary screen on YEPD + G418. Incubate for 2 days at 30°C.

Fig. 4. Identifying regulators of a G1-phase cell cycle promoter element using a HIS3 reporter. (a) Modified R-SGA screening approach. In an R-SGA screen designed to identify inhibitors of SBF-dependent transcription, a cln3Δhis3Δ query strain bearing an integrated copy of the HIS3 reporter gene under the control of four consensus Swi4/Swi6-dependent cell cycle box (SCB) elements was crossed to the yeast gene-deletion array (see Note 9) (20). This query strain is a histidine auxotroph as a result of the absence of CLN3, which is required for expression of the SCB-HIS3 reporter. MATa xxxΔ::kanrcln3Δhis3Δ strains containing the integrated reporter were isolated using a variation of the R-SGA selection steps described. (b) Detection of SCB-HIS3 repressors. MATa xxxΔ::kanrcln3Δhis3Δ strains containing the SCB-HIS3 reporter were pinned onto SD −His + 30 mM 3-AT (high concentration) to identify strains that express sufficient levels of HIS3 to confer viability. These are deletion mutants in which the transcriptional defect resulting from cln3Δ is suppressed. WHI5 (boxed) was identified as a repressor of SCB-HIS3. Reproduced from ref. 20 with permission from Elsevier Science. (c) Confirmation of SCB-HIS3 repression by WHI5. Serial dilution growth assays confirmed the ability of the cln3Δhis3Δwhi5Δ strain containing the SCB-HIS3 reporter to grow on SD −His + 30 mM 3-AT.

Reporter-Based Synthetic Genetic Array Analysis

a HIS3

xxxΔ

X

cln3Δ his3Δ MATα

MATa

Modified R-SGA selection steps

HIS3

cln3Δhis3Δ xxxΔ b

whi5Δ

SD-His + 30 mM 3-AT c

WT

cln3Δ

cln3Δwhi5Δ SD-His + 30 mM 3-AT

SD-Ura

67

68

Sassi et al.

2. Introduce the CCpr-HIS3 plasmid into the deletion strains by standard yeast transformation, and plate onto SD −Ura. Incubate for 2–3 days at 30°C. 3. Aliquot 200 μl of SD −Ura liquid medium into each well of a 96-well round bottom plate. 4. Inoculate each candidate regulator gene-deletion strain carrying the CCpr-HIS3 plasmid into the 96-well round bottom plate in duplicate. 5. Grow the cultures for 2 days at 30°C. Mix the cultures by pipetting after the first day. 6. Inoculate 5 μl of each culture into 315 μl of fresh SD −Ura in a 96-well round bottom plate. 7. Perform five 1:5 serial dilutions of each culture in the same 96-well plate (30 μl added to 120 μl of fresh SD −Ura liquid medium each). 8. Spot 4 μl of all dilutions of each strain onto SD −Ura, SD −His + 10 mM 3-AT and SD −His + 75 mM 3-AT plates. Incubate the plates for 2 days at 30°C. 9. Repeat steps 2–8 using the Controlpr-HIS3 plasmid. 10. Score strain growth on the SD −His + 3-AT plates compared to the SD −Ura control plate for both the CCpr-HIS3 and Controlpr-HIS3 plasmids to identify high-confidence regulators that are CCpr-specific (see Note 21). Fig. 4c shows an example of serial dilution growth assays performed with xxxΔ::kanr strains containing a HIS3 reporter.

4. Notes 1. The antibiotic G418 is added to YEPD and to enriched sporulation plates (at a lower concentration) to minimize contamination. 2. Ammonium sulfate interferes with the sensitivity of cells to G418 (21). We therefore use MSG as the nitrogen source in our synthetic plates containing G418. 3. In addition to 2-µm reporter constructs, we have used CEN-based constructs and integrated CCpr reporters in our R-SGA screens. The reporter context will depend on the promoter under study. Weak promoters may require the use of the 2-µm origin of replication to sufficiently amplify the reporter signal for robust detection, whereas CEN-based or integrated CCpr reporters are generally more appropriate for strong promoters.

Reporter-Based Synthetic Genetic Array Analysis

69

4. For the lacZ construct, promoters can be cloned as XhoI–XhoI or SalI–SalI fragments. For pHIS3, promoters can be cloned as EcoRI–BamHI fragments. 5. In addition to lacZ and HIS3 reporters, we are currently expanding the R-SGA system to include green fluorescent protein (GFP) and derivatives of red fluorescent protein (RFP) reporters. 6. RPL39 encodes a ribosomal protein, and ACT1 encodes actin. Expression of these genes is constitutive and independent of cell cycle regulation (2). In our screens, expression of reporters under the control of RPL39 or ACT1 promoters is unaffected in the vast majority of mutants defective in known cell cycle transcription factors. 7. We maintain a master copy of the yeast gene-deletion set in 384-array format on YEPD + G418 plates at 4°C. The plates are propagated as needed in 384-format for maintenance and 768-format for screens. Each colony is represented in duplicate in the 768-format, and our current array is 14 plates. Yeast colonies adjacent to empty spots and those on the borders of each plate in a 768-density array have access to more nutrients and generally grow faster and larger than neighboring colonies that are completely surrounded by others. Reporter activity in these larger colonies can be higher than in genetically identical smaller colonies positioned in highdensity areas. Steps can be taken to minimize these positional effects: (1) each plate in the 384-density master array can be designed with borders consisting of a neutral strain carrying all the markers required for the R-SGA selection steps, (2) the neutral strain can be positioned at fixed intervals within the array as a reference for wild-type reporter levels, (3) for a given CCpr, known regulators can be positioned at fixed intervals within the array as references for enhanced or reduced reporter activity. 8. Using the R-SGA protocol, deletion strains containing CCprreporter plasmids can be recovered 16 days from the generation of the plasmid-bearing starting strain. Approximately 2 h of pinning time is required to replicate one complete yeast gene-deletion set at each selection step. Recently, we have made use of the Singer RoToR HDA platform (Singer Instruments), which cuts pinning time down to approximately 20 min for each selection step per screen. 9. We have screened the yeast overexpression array (4) using query strains bearing integrated, nourseothricin (cloNAT) resistance-marked cell cycle reporters, and are currently making adaptations to screen essential genes using the Tet-promoters Hughes collection (yTHC) (22). We used a variation

70

Sassi et al.

of the R-SGA method that incorporated a cln3Δ allele into the starting strain to identify Whi5 as an inhibitor of G1/S transcription (Fig. 4) (20). One could modify the reporter constructs to include complete ORFs, facilitating the identification of regulators of cell cycle protein expression and/or stability. Further, specific promoter elements rather than full promoter sequences could be screened: we have screened both Swi4/6-dependent cell cycle box (SCB) (Fig. 4) and MluI cell cycle box (MCB) (Fig. 2) promoter elements using R-SGA. 10. To successfully cross query strains to the array, lawns must be as dense and as even as possible. Two or three lawns are sufficient for mating to the entire yeast gene-deletion array. 11. Spores are pinned in duplicate onto two separate SD −Ura/ Leu/Arg/Lys + can/thia plates, effectively generating two copies, or sets. Each set is carried through the remaining haploid selection steps such that four colony replicates of each deletion mutant (two in each set) are ultimately analyzed for reporter activity. 12. Care must be taken when pouring the X-gal top agarose mixture, as excessive force can wash the colonies away. Using 0.5% agarose, solidification of the X-gal solution usually occurs within 30 s. 13. Color development can take anywhere from 15 min to overnight depending on the strength of the promoter and the concentration of X-gal. The array-based β-galactosidase assay conditions described above have been optimized for promoters screened in our laboratory. Prior to assaying lacZ reporter activity on a genome-wide scale, it is important to perform a small-scale overlay experiment to determine the optimum assay conditions (concentration of X-gal, incubation time and temperature) and examine basal reporter activity. Pilot experiments should include several transformants and should be done using pinned colonies to mimic the final assay conditions. 14. Under ideal experimental conditions, colonies with wildtype levels of lacZ expression will appear light blue, those with reduced expression will appear white, and those with enhanced expression will appear dark blue. In practice, a range of color intensities is usually observed. It is helpful to use known regulators as benchmarks with which to develop scoring criteria to identify new candidates. We have found the identification of activators significantly easier than the identification of repressors using lacZ reporters; the ability to identify repressors of a given CCpr will largely depend on the basal activity of the reporter and may not be possible with certain promoters.

Reporter-Based Synthetic Genetic Array Analysis

71

15. Plates covered with X-gal top agarose can show uneven distribution of blue color, particularly around the edges. The confirmation plates containing mutant strains next to wild type are therefore especially important. 16. In our screens, non-cell-cycle-specific regulators generally include genes with roles in general transcription, including components of the RNA polymerase II mediator complex (MED1, NUT genes, SRB genes). We also find genes involved in transcriptional elongation, histone modification, chromatin remodeling, and mRNA stability, likely as a result of their roles in the general regulation of gene expression. Genes that function in cellular respiration (HAP genes, COX genes), protein translation, trafficking, and degradation are also frequently identified, although the reasoning behind their regulatory effects is less clear. 17. The number of candidate regulators identified in each screen is variable and depends on the promoter under study and the scoring criteria. In a CLN2pr-lacZ R-SGA screen, approximately 500 candidate regulators were identified in the primary screen and 235 were confirmed in miniarray format (N. Bastajian, unpublished data). Candidate regulators are not limited to those that bind the promoter of interest; R-SGA has the capacity to identify both direct and indirect regulators. Care should therefore be taken when inferring regulatory relationships. 18. The sensitivity of detecting reporter-based transcriptional effects from whole colonies in array format is dependent on the promoter, the reporter, and the scoring system. In a CLN2pr- lacZ screen, we identified known regulators of CLN2 expression, including Swi4, Swi6, and Bck2 (N. Bastajian, unpublished data). 19. In some cases, CCpr-reporter constructs are somewhat leaky, allowing survival on SD −His even in the absence of activated transcription, probably as a result of basal transcription. To overcome this problem, the competitive inhibitor of the imidazoleglycerol-phosphate dehydratase enzyme encoded by HIS3, 3-AT, is added to the plates. The level of background expression varies depending on the reporter construct and genetic background. It is therefore essential that prior to every screen, serial dilution growth assays (see Subheadings 2.4.2 and 3.4.2) be performed to determine the minimum amount of 3-AT that effectively inhibits growth of the starting strain. Generally, this will lie in the range of 10–100 mM, but in certain cases it may be higher or lower. The advantage of employing this inhibitor is that for any given strain, one can identify genes that both upregulate and downregulate expression of the reporter gene by analyzing

72

Sassi et al.

growth at concentrations both above and below the minimum inhibitory concentration. 20. The basis for comparison to the final R-SGA selection plates is that any xxxΔ::kanr strain that is inherently slow growing will produce smaller than average colonies on both the final selection and SD −His + 10 mM 3-AT plates. These strains can then be removed from the list of putative activators. In addition to visual inspection of colony size, we use a computer-based scoring system developed in the Boone Laboratory. 21. In the HIS3 reporter assays, subtle differences in strain survival observed upon pinning are generally obvious using serial dilution growth assays.

Acknowledgments We thank Michael Costanzo and Jonathan Millman for sharing unpublished results. We are grateful to Helena Friesen for helpful discussions and critical reading of the manuscript. This work was supported by grants to B.A. from the Canadian Institutes for Health Research, the National Cancer Institute of Canada, and Genome Canada through the Ontario Genomics Institute.

References 1. Cho, R. J., Campbell, M. J., Winzeler, E. A., et al. (1998). A genome-wide transcriptional analysis of the mitotic cell cycle. Mol. Cell. 2, 65–73. 2. Spellman, P. T., Sherlock, G., Zhang, M. Q., et al. (1998). Comprehensive identification of cell cycle-regulated genes of the yeast Saccharomyces cerevisiae by microarray hybridization. Mol. Biol. Cell. 9, 3273–3297. 3. Pramila, T., Wu, W., Miles, S., et al. (2006). The Forkhead transcription factor Hcm1 regulates chromosome segregation genes and fills the S-phase gap in the transcriptional circuitry of the cell cycle. Genes Dev. 20, 2266–2278. 4. Sopko, R., Huang, D., Preston, N., et al. (2006). Mapping pathways and phenotypes by systematic gene overexpression. Mol. Cell. 21, 319–330. 5. Iyer, V. R., Horak, C. E., Scafe, C. S., et al. (2001). Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF. Nature 409, 533–538.

6. Simon, I., Barnett, J., Hannett, N., et al. (2001). Serial regulation of transcriptional regulators in the yeast cell cycle. Cell 106, 697–708. 7. Horak, C. E., Luscombe, N. M., Qian, J., et al. (2002). Complex transcriptional circuitry at the G1/S transition in Saccharomyces cerevisiae. Genes Dev. 16, 3017–3033. 8. Lee, T. I., Rinaldi, N. J., Robert, F., et al. (2002). Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 298, 799–804. 9. Ihmels, J., Friedlander, G., Bergmann, S., et al. (2002). Revealing modular organization in the yeast transcriptional network. Nat. Genet. 31, 370–377. 10. Chen, H. C., Lee, H. C., Lin, T. Y., et al. (2004). Quantitative characterization of the transcriptional regulatory network in the yeast cell cycle. Bioinformatics 20, 1914–1927. 11. Winzeler, E. A., Shoemaker, D. D., Astromoff, A., et al. (1999). Functional characterization

Reporter-Based Synthetic Genetic Array Analysis

12.

13.

14.

15.

16.

of the S. cerevisiae genome by gene deletion and parallel analysis. Science 285, 901–906. Tong, A. H., Evangelista, M., Parsons, A. B., et al. (2001). Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294, 2364–2368. Giaever, G., Chu, A. M., Ni, L., et al. (2002). Functional profiling of the Saccharomyces cerevisiae genome. Nature 418, 387–391. Tong, A. H., Boone, C. (2005). Synthetic genetic array analysis in Saccharomyces cerevisiae. In Yeast Protocols, 2nd edn., Methods in molecular biology, vol. 313 (Xiao, W., ed.).. Humana, Totowa, NJ, pp. 171–192. Tong, A. H., Boone, C. (2007). Highthroughput strain construction and systematic synthetic lethal screening in Saccharomyces cerevisiae. In Yeast Gene Analysis, 2nd edn., Methods in microbiology, vol. 36 (Stansfield, I. and Stark, M., eds.).. Elsevier, Burlington, MA, pp. 369–386; 706–707. Johnson, A. D., Herskowitz, I. (1985). A repressor (MATa2 product). and its operator control expression of a set of cell type specific genes in yeast. Cell 42, 237–247.

73

17. Christianson, T. W., Sikorski, R. S., Dante, M., et al. (1992). Multifunctional yeast high-copynumber shuttle vectors. Gene 110, 119–122. 18. Brachmann, C. B., Davies, A., Cost, G. J., et al. (1998). Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast 14, 115–132. 19. Verma, R., Smiley, J., Andrews, B., Campbell, J. L. (1992). Regulation of the yeast DNA replication genes through the Mlu I cell cycle box is dependent on SWI6. Proc. Natl. Acad. Sci. U S A. 89, 9479–9483. 20. Costanzo, M., Nishikawa, J. L., Tang, X., et al. (2004). CDK activity antagonizes Whi5, an inhibitor of G1/S transcription in yeast. Cell 117, 899–913. 21. Cheng, T. H., Chang, C. R., Joy, P., et al. (2000). Controlling gene expression in yeast by inducible site-specific recombination. Nucl. Acids Res. 28, E108. 22. Mnaimneh, S., Davierwala, A. P., Haynes, J., et al. (2004). Exploration of essential gene functions via titratable promoter alleles. Cell 118, 31–44.

Chapter 5 The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy Heimo Wolinski, Klaus Natter, and Sepp D. Kohlwein Summary Despite its small size of 5–8 mm – only one order of magnitude above the wavelength of visible light – yeast has developed into an attractive system for light microscopic analysis. First, the ease of genetic manipulation and integrative transformation have opened numerous experimental strategies for genome-wide tagging approaches, e.g., with fluorescent proteins (as discussed in several chapters of this issue). Second, the large number of cells that can be simultaneously visualized provides an excellent basis for statistical image analysis, resulting in reliable morphological or localization information. Third, the flexibility of yeast cultivation in terms of biochemical manipulation, rapid cellular growth, mutant isolation or drug susceptibility offers an unprecedented spectrum of possibilities for in vivo functional studies, and analysis of cellular dynamics and organelle inheritance. Although yeast in itself is an interesting cellular system, its “prototype character” in understanding cellular metabolism, physiology, and signaling in eukaryotes accounts for its popular use in technology development and biomedical research. Here we discuss experimental strategies for live yeast cell imaging, geared towards imaging-based large-scale screens. Major emphasis is on the methods for immobilizing cells under “physiological” conditions, with minimum impact on yeast. We also point out potential pitfalls resulting from live cell imaging that once again stresses the necessity for extremely careful experimental design and interpretation of data resulting from imaging experiments. It goes without saying that these problems are not restricted to yeast and are also highly relevant to “large” cells. If an image tells more than a thousand (perhaps misleading?) words, the ease of obtaining “images” thus rather suggests analyzing many thousands of images, to come up with one relevant and biologically significant conclusion. Key words: Yeast, Saccharomyces cerevisiae, Fluorescence microscopy, Green fluorescent protein, High-content imaging, Large-scale microscopy, Live cell microscopy

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_5, © Humana Press, a part of Springer Science + Business Media, LLC 2009

75

76

Wolinski, Natter, and Kohlwein

1. Introduction The ease of obtaining “images” in microscopy requires consideration of numerous experimental requirements and constraints in order to obtain reliable and meaningful results. This involves: (a) construction of cells expressing fluorescent protein (FP) fusions, (b) cell cultivation prior to microscopy, (c) cell labeling with vital dyes, (d) preparation and maintenance of cells during microscopy, (e) microscope setup and imaging, (f) downstream processing of images. The emphasis of the methods described below was put on generating (a) a rather simple and reliable setup for standard live yeast cell imaging and imaging-based screens that are (b) also applicable without dedicated imaging systems on “standard” upright and inverted confocal microscopes or other types of high-resolution imaging devices. The advantage of looking at many – 100s, 1,000s – of yeast cells at once provides a significant advantage over larger cells, to obtain reliable localization patterns. However, It becomes immediately evident that yeast cell populations are never homogeneous as a result of replicative or chronological aging, different stages of the cell cycle etc. This may not necessarily impact interpretation of global localization patterns in most cases, but obviously may affect the interpretation of localization patterns that are changing during the cell cycle or the aging process. Observation of such fluctuations of protein localization patterns obviously yields very important functional information. However, there is also the need to clearly discriminate “physiological” patterns from induced deficiencies during microscopy, which requires very careful control of the microscopy setup. In some examples we will highlight – unexpected – observations for rapidly induced changes in organelle morphology or protein localization due to changes in the microenvironment during microscopy. Labeling of cells with vital dyes is particularly error-prone as cellular age and physiology may significantly affect uptake of dyes and labeling efficiency, e.g., through the activity of pleiotropic drug resistance pumps that may take care, more or less efficiently, of xenobiotic vital dyes. For instance, elimination of the Pdr1 transcriptional regulator yields drastically stimulated uptake and staining intensity with Nile Red, a popular lipophilic dye used to detect lipidic structures in yeast (1). Additionally, the illumination of the specimen itself may have a detrimental effect on the physiology and cellular morphology. Thus, reducing the intensity and exposure time to (laser) light is of utmost importance in live cell imaging, representing challenges for efficient labeling or detectable levels of expression, and optimized sensitivity of the microscopy equipment (see Note 1).

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

77

Preparation problems multiply for cell handling in “highthroughput” setups, i.e., for imaging-based screens. In such experiments, mutants may be identified with altered organelle morphology or subcellular distribution of a particular FP-tagged “query” construct that defines a specific subcellular structure. Thus, factors that are important for organelle morphology, protein trafficking, protein expression and localization may be identified that may not be accessible with any other type of screen. Vital dyes appear as an easy solution to analyze mutant collections, but are obviously restricted to organelles that can be specifically labeled; further considerations involve the impact of dye uptake and labeling efficiency in a particular mutant background that may obscure the imaging pattern. Alternatively, FP-tagged query constructs expressed from plasmids can be rather easily introduced into the mutant collections, by mass transformation or cytoduction (R. Rothstein, personal communication). Main disadvantages of plasmid-borne constructs, however, are the requirement for selective conditions for plasmid maintenance, limiting its practical application, and, more severely, a rather inhomogeneous expression of plasmidborne FP fusions in the cell population; this inhomogeneity in fluorescence intensity may obscure subtle changes in localization patterns, thus leading to false-positive and false-negative results. Also, such labeling patterns are not amenable to computer-based statistical image analysis (see Note 2). The large number of cell types – in the range of 6,000 – that are to be imaged for mutant screens requires some compromising, in particular at the level of cell preparation. The time requirements for imaging challenges the logistics for cell handling in order to analyze cell populations that are cultivated for a comparable time window (see Note 3). A critical issue involves the maintenance of “physiological” conditions during the microscopy. A rather simple agar-based method for cell immobilization and maintenance is described in detail below. This method allows for preparation of slides with typically 96 (up to 2× 96) colonies on standard microscope slides, both for upright and inverse microscopes. The spacing of colonies is compatible with the 1,536 colony format on standard rectangular plates that can be ideally and precisely handled by the Singer Rotor replicating robot. A few examples with “critical” proteins or subcellular structures that obviously react highly sensitively to the microenvironment on the slide may illustrate the reliability of this immobilization method. It must be stressed, however, that conditions under the microscope may never completely represent conditions in liquid culture or on the surface of solid media plates. The selection of the correct microscope settings obviously depends on the signal intensity of the fluorescent sample (see Note 4).

78

Wolinski, Natter, and Kohlwein

The optical path in general and the objectives in particular are major determinants of the resolution that can be achieved. For yeast imaging typically oil immersion objectives with high numerical apertures (NA = 1.4) are used. For optimum resolution, thus, the mounting medium of the specimen should match the refractive index of the immersion fluid (see Note 5). The downstream processing of images obviously depends on the specific experimental problem; typically, selection of images occurs “by eye,” but more standardized procedures for microscopic image acquisition need to be considered, i.e., for the evaluation of the dynamic range, saturation, and contrast of the digital image (see Note 6). Only recently, methods have been implemented for quantitative feature extraction of yeast fluorescence images (2–4). Repositories for image data are available through the Saccharomyces Genome Database (http://www.yeastgenome.org/) at: YeastGFP database (http://yeastgfp.ucsf.edu) (5); Yeast Protein Localization database (http://YPL.uni-graz.at) (6, 7); Organelle.db (http://organelleview.lsi.umich.edu) (8); TRIPLES (http://ygac.med.yale.edu/triples/) (9). Morphological data of yeast mutants, rather than localization data, are available at the Saccharomyces Cerevisiae Morphological Database (http://yeast.gi.k.u-tokyo.ac.jp/) (10).

2. Materials 2.1. Yeast Strains

Wild type strains BY4741 (MATa his3D1 leu2D0 met15D0 ura3D0) and BY4742 (MATa his3D leu2D0 lys2D0 ura3D0) and strains Acc1-GFP, CoxIV-GFP, and Elo3-GFP harboring chromosomally integrated GFP fusion constructs are obtained from Euroscarf (Institute of Microbiology, Johann Wolfgang GoetheUniversity Frankfurt, Germany) as well as from Invitrogen, Inc. (5). Strains Y7092 (MATa can1D::STE2pr-Sp_his5 lyp1D his3D1 leu2D0 ura3D0 met15D0) and Y8205 (MATa can1D::STE2pr-Sp_his5 lyp1D::STE3pr-LEU2 his3D1 leu2D0 ura3D0) used for construction of query strains were provided by Charlie Boone (Donnelly Centre for Cellular and Biochemical Research, University of Toronto). Template plasmids for fluorescence protein fusions can be obtained through NCRR Yeast Resource Center, Seattle (http://depts.washington.edu/yeastrc/pages/micro.html) or from the authors.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

2.2. Cell Culture

79

1. Bacto™ yeast extract. 2. D(+)-glucose monohydrate. 3. Bacto™ peptone. 4. Bacto™ agar. For preparation of yeast media, yeast extract (1% w/v) and peptone (2% w/v) are weighed in an appropriate flask and filled up with distilled water to 9/10 of the final volume. Autoclave at 121°C for 20 min; add 1/10 volume of 10× glucose solution (2% w/v final concentration) by sterile filtration to the media, mix well and dispense into sterile flasks or glass tubes. For solid media plates 2% Bacto™ agar is added to the solution prior to autoclaving. For preparation of agar plates and agar-coated slides without nutrients 2% (w/v) Bacto™ agar in distilled water is autoclaved at 121°C for 15 min. For small scale cultivation and vital staining, cells are cultivated in 6- or 12-well plates in 5–2 ml of the medium under shaking and temperature control (30°C) on a thermomixer. Plates for selection of positive transformants contain 200 mg/l G418 sulfate (Sigma-Aldrich, Inc.) or 200 mg/L nurseothricin (Werner BioAgents, Inc.). Media for the arraybased introduction of chromosomally integrated GFP fusions into the deletion mutant collection are prepared according to Tong and Boone (11).

2.3. Fluorescence Labeling

1. MitoTracker® Red CM-H2XRos (Invitrogen, Inc., Cat.-No. M7513). Prepare stock solution of 1 mg/ml MitoTracker® Red CMH2XRos in dimethylsulfoxide (DMSO). Store 20 ml aliquots of the stock solution in 500 ml reaction tubes at −20°C. One microliter of MitoTracker® Red CM-H2XRos was added to 1 ml of a cell suspension in a 1.5-ml reaction tube (final concentration: 1 mg/ml). Labeling was performed for 10 min at RT without subsequent washing of cells. 2. DASPMI (4-(4-(dimethylamino)styryl)-N-methylpyridinium iodide (Invitrogen, Inc., Cat.-No D288): For preparation of stock solution and labeling see MitoTracker Red. 3. FM4–64 (12) (Invitrogen, Inc., Cat.-No. T13320). For preparation of stock solution and labeling see MitoTracker Red. Labeling was performed for 60 min at RT. 4. Nile Red (Invitrogen, Inc., Cat.-No. N1142). For preparation of stock solution and labeling see MitoTracker® Red. Centrifuge Nile Red stock solution briefly in a mini centrifuge prior to cell staining, to remove precipitates that may have formed. 5. Formaldehyde solution min. 37% (Merck, Inc.).

80

Wolinski, Natter, and Kohlwein

2.4. Cell Preparation

1. Singer RoToR High Density Array robot (Singer Instrument Co.Ltd). 2. Cultivation plates (Plus plates; Singer Instrument Co. Ltd). 3. Plastic string or metal wire (0.2 mm diameter). 4. 4× magnifying glass with stand and illumination feature. 5. Scalpel No. 10 (Martor Inc., Germany). 6. Low-melting agarose.

2.5. Microscopy Hardware and Imaging Software

1. Microscope slides 76 × 26 mm. 2. Coverslips 50 × 24 mm #1. 3. Leica TCS SP2 confocal microscope with differential interference contrast (DIC) optics, acousto optical tunable filter (AOTF), acousto optical beam splitter (AOBS); spectral detection. Märzhäuser MCX-2 stage controller (Märzhäuser, Inc.). Leica ScanWare™ and Leica TCS SP2 microscope control and imaging software v2.60.1537. 4. Leica TCS4d confocal microscope with DIC optics, AOTF. 5. Laser sources: 488 nm (argon-laser; GFP, DASPMI excitation), 543 nm (green helium-neon laser; MitoTracker® Red CM-H2XRos, Nile Red excitation); ArKr laser (488 nm; 543 nm). 6. Leica objectives: HCX PL APO 100×/1.4–0.7 OIL CS; HCX PL APO 63×/1.32–0.6 OIL CS; HCX PL APO 63×/1.3 GLYC CORR, Leica HCX PL APO 63×/1.2 W CORR. 7. Bioptechs Objective Heater System (Bioptechs, Inc.). 8. Adobe Photoshop™ CS (Adobe, Inc., USA). Image J, public domain image processing software (author: Wayne Rasband, National Institute of Mental Health, Bethesda, Maryland, USA). amira™ 4.0 (Mercury Computer Systems, Inc.).

3. Methods 3.1. Introduction of Chromosomally Tagged GFP Fusion into the Deletion Mutant Collection

GFP fused genes are introduced into a mutant collection according to the “synthetic genetic array (SGA)” protocol, as described by Tong et al. for the systematic construction of double mutant strains (11, 13). GFP fusions in appropriate query strains suitable for the SGA protocol are constructed as follows (Fig. 1):

3.1.1. Construction of the GFP Query Strain

The GFP cassette is integrated into the query strain background required for the SGA protocol, harboring the following markers: MATa can1D::STE2pr-Sp_his5 lyp1D his3D1 leu2D0 ura3D0

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

81

Fig. 1. Schematic representation of introducing a chromosomally integrated GFPlabeled query construct into the deletion mutant collection. (Adapted by permission from Macmillan Publishers Ltd (ref. 14), copyright 2007.) This method is applicable for both N- or C-terminally tagged variants, and relies on the basic protocol by Tong and Boone (11,13 ).

met15D0. If the integrated GFP-cassette linked with the ClonNAT marker is already available in a different background, the query strain is generated by mating with a switching strain harboring the required markers, and subsequent sporulation and selection of haploid progeny with the correct genotype. A switching

82

Wolinski, Natter, and Kohlwein

strain obtained from Charlie Boone (Donnelly Centre for Cellular and Biochemical Research, University of Toronto), Y8205 MATa can1D::STE2pr-Sp_his5 lyp1D::STE3pr-LEU2 his3D1 leu2D0 ura3D0 allows for selection of suitable query strains on selective media: 1. Mate GFP-expressing strain (MATa) with strain Y8205. 2. Pick zygotes. 3. Sporulate on acetate plates. 4. Select a-haploid progeny on leucine-free plates containing thialysine, canavanine. 5. Select for a query strain on leucine-free plates that contain thialysine, canavanine, nurseothricin. 6. Test for GFP expression (microscopy) and correct integration of markers by PCR. If the integrated GFP-cassette is marked with the kanMX resistance marker instead of nurseothricin resistance (clonNAT), an appropriate query strain can be obtained by marker switching (11) (see Note 7). The PCR product is transformed into the recipient strains by standard lithium acetate transformation. Construction of C-Terminal GFP Fusions (see Note 8)

The primers for amplification of the GFP-ClonNAT cassette contain 50 bases homologous to the 3¢-region of the gene of interest. The template plasmid is pKN082 (Natter, unpublished). Sequences for binding on the vector template are 5¢-GTGAGCAAGGGCGAGGAG-3¢ and 5¢-CAGTATAGCGACCAGCATTCAC-3¢. 1. Amplify GFP-ClonNAT cassette using appropriate primers for insertion 3¢ to the reading frame. 2. Transform PCR product into the SGA-compatible query strain (MATa can1D::STE2pr-Sp_his5 lyp1D his3D1 leu2D0 ura3D0 met15D0). 3. Select transformants on plates containing nurseothricin. 4. Test for GFP expression (microscopy) and correct integration of markers by PCR.

Construction of N-Terminal Fusions

Construction of N-terminally GFP-tagged proteins that are expressed under endogenous promoter control requires a twostep cloning strategy. We prefer to use the endogenous promoter for reasons discussed in the introduction. We constructed a vector, pCGCNm, bearing the ClonNAT cassette followed by a multiple cloning site and monomeric GFP (Natter, unpublished). The promoter region of the gene of interest is cloned into the MCS of pCGCNm. The entire cassette, NATR-Promoter-mGFP, is amplified with primers containing 50 bases, homologous regions to the 5¢-region of the target gene and the annealing sequences

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

83

5¢-ACATGGAGGCCCAGAATACCC-3¢ (fwd) and 5¢-CTTGTACAGCTCGTCCATGCC-3¢ (rev), respectively. 1. Amplify the promoter sequence of the gene of interest, using flanking primers with compatible restriction sites for insertion into the MCS of plasmid pCGCNm. 2. Amplify ClonNAT-gene promoter-GFP-cassette using appropriate primers for insertion 5¢ to the reading frame. 3. Transform PCR product into the SGA-compatible query strain (MATa can1D::STE2pr-Sp_his5 lyp1D his3D1 leu2D0 ura3D0 met15D0). 4. Select transformants on plates containing nurseothricin. 5. Test for GFP expression (microscopy) and correct integration of markers by PCR. 3.2. Cell Cultivation (see Notes9 and 10)

1. Replicate mutant cells on suitable agar media plates (Plus plates, Singer Instrument Co. Ltd.) to a density of 1,536 colonies using a Singer Rotor HDA robot. 2. Grow cells for 24 h at 30°C. 3. Replicate cells to fresh plates. 4. Repeat replication step to generate very small colonies (“dilution”), on desired media plates (see Note 10).

3.3. Immobilization of Cells for Live Cell Imaging 3.3.1. Setup for Vital Dye Labeling

1. Boil 2.5% low melting temperature agarose in aqua dest. or in appropriate growth media in a microwave oven (600 W) until the agarose solution is completely clear. Alternatively, use 2% Bacto™ agar instead of agarose (see Note 11). 2. Transfer 5 ml of melted agarose solution into a 15-ml plastic tube and equilibrate at ~50°C in a heated water bath. 3. Add 5 ml of the fluorescence dye from stock solution (final concentration typically 1 mg/ml) and vortex for 30 s. (a) Optionally add 135 ml formaldehyde solution (37% in water; final concentration 1% v/v) to the agar solution (Note: Addition of a low concentration of formaldehyde kills the cells and enables more homogeneous labeling with vital dyes due to inactivation of drug pumps. Cell morphology as imaged by DIC is not compromised in these preparations). 4. Place a standard microscope slide (76 × 26 mm) on a flat surface and pipette 3 ml of the agarose solution on the microscope slide until it is completely covered. 5. Allow gelation of the agarose on the slide for 15 min at RT. 6. Add 1 ml of (labeled) yeast cell suspension to the center of the agarose layer. Mount the cell preparation with a large (50 × 24 mm) coverslip (see Note 12).

84

Wolinski, Natter, and Kohlwein

3.3.2. Setup of Yeast Cell Arrays for Large-scale High-content Imaging (Fig. 2)

1. Put a sheet of plastic (2 mm) as solid support for the agarlayer into a Plus plate. 2. Pour agar solution (cooled to ~50°C) into the plate. 3. Allow gelation of the agar for 15 min at RT. 4. Replicate cell colonies in the 1,536 format with a Singer Rotor replicator. 5. Remove agar block and dissect into 16 pieces with 96-colonies each, preferably by cutting with a thin plastic string or metal wire.

Fig. 2. Preparation of yeast cell-arrays for large-scale, high-content microscopy. 1, Put a sheet of plastic (2 mm) as solid support for the agar-layer into a Plus plate. 2, Pour 50 ml agar solution (cooled to ~50°C) into the plate (~5 mm thickness). 3, Replicate cell colonies in the 1,536 format with a Singer Rotor replicator. 4/5, Remove agar block together with solid support. 6, Dissect agar block into 16 pieces with 96 colonies each. 7, Mount agar block with 96 colonies onto standard glass slide with an alignment scheme (printed with a laser printer on transparency film), for reproducible positioning on the microscope. 8, Cut horizontal and vertical slits between the yeast colonies prior to covering colonies with a cover slip. This prevents air bubbles and mixing of colonies.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

85

6. Cut horizontal and vertical slits separating the colonies. 7. Mount agar block with 96 colonies on standard glass slide with the aid of an alignment scheme (printed with a laser printer on transparency film), for easier and reproducible positioning on the microscope stage. 8. Cover colonies with a large (50 × 24 mm) cover slip. 9. Press the coverslip carefully onto the agar surface to remove air bubbles. 10. Perform imaging, using manual or automatic positioning, software-based autofocus and multiple section image acquisition. 11. Carry out image storage, data processing and subsequent off-line image evaluation (see Note 13). 3.4. Microscopy Setup 3.4.1. Transmission Microscopy: Differential Interference Contrast

1. Pipet ~1 ml of cell suspension on standard microscope slides or on agarose-coated slides. 2. Cover cells with a large coverslip and mount slides under the microscope; set temperature controller of the objective heater to the desired temperature. 3. Visualize cells using DIC optics. Note that the condenser position needs to be adjusted for cells mounted on agarosecovered slides, due to the extended distance, for optimized illumination of the specimen.

3.4.2. Fluorescence Microscopy (see Note 14)

1. Pipet ~1 ml of cell suspension on standard microscope slides or on agarose-coated slides. 2. Cover cells with a large coverslip and mount slides under the microscope; set temperature controller of the objective heater to the desired temperature. 3. Optimize laser intensity, fluorescence filter (fluorescence emission range) and photomultiplier settings, using the lowest possible laser intensity. 4. Allow the cells to adjust to the conditions under the microscope (see Note 15). 5. Re-adjust focus and image cells in areas not previously exposed to the laser light. This experimental setup allows reliable 3D imaging of doublelabeled cells (ELO3-GFP, MitoTracker) over several hours (Fig. 3).

3.5. Specific Applications: FRAP and FLIP (Fig. 4)

Laser scanning microscopy allows for application of specific techniques, such as fluorescence recovery after photobleaching (FRAP) or fluorescence loss in photobleaching (FLIP), to investigate organelle and protein dynamics. Modern microscopes provide features to restrict laser illumination to regions of interest in the

86

Wolinski, Natter, and Kohlwein

Fig. 3. Four-dimensional imaging of multi-labeled yeast cells during cellular growth. Endoplasmic reticulum was labeled with Elo3-GFP (green). Mitochondrial structures were labeled with MitoTracker® Red CM-H2XRos (red). Cells were cultivated on agarose/complete media covered slides. Selected single optical sections (fluorescence and transmission images) of the acquired 4D data set (25 optical sections each) are shown. The image at the bottom right represents a polygon-based 3D reconstruction (longitudinal cut through the surface model) of the masked cell shown in the top panel at right. Bar, 10 mm (see Color Plates).

nanometer scale, which are required for bleaching subcellular yeast structures at the limit of optical resolution. (see Note 16). 1. Precultivate GFP-fusion expressing cells in appropriate media under vigorous shaking at 30°C. 2. Concentrate cells by centrifugation and mount ~1 ml of cell suspension on standard glass slides; cover with a large coverslip. 3. Define scanning area with low intensity laser light. Include control cells within the field of view that are not going to be bleached. 4. Define region of interest (ROI) for bleaching. Optionally define a ROI for detection of fluorescence intensity outside the bleaching ROI (FLIP). 5. Acquire a series of images before bleaching using fast scanning mode. 6. Apply short bleach pulse (FRAP, FLIP) or longer bleach pulses (FLIP). 7. Acquire a series of post bleach images using the same settings as applied for prebleach scanning.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

87

8. Normalize experimental data (background subtraction, correction for photobleaching during measurement). 9. Estimate quantitative parameters (mobile and immobile fraction, recovery half-time, diffusion coefficient).) 3.6. Potential Pitfalls of Live Cell Fluorescence Microscopy (see Note 17)

The popular use of some fluorescence dyes has prompted us to discuss potential pitfalls resulting from inappropriate preparation or imaging techniques. Examples illustrate the impact of imaging conditions on organelle morphology or protein localization. Thus, various settings need to be tested to obtain reliable imaging information; the following scenarios may aid in designing appropriate control setups.

3.6.1. Fluorescence Activation of DASPMI (Fig. 5; See Note 18)

1. Label cells grown in complete media to stationary growth phase for 10 min in suspension with 1 mg/ml DASPMI and mount on agarose-coated slides. 2. Set (increase) zoom factor to obtain a field of view (image) of ~40 × 40 mm (100× magnification objective). 3. Image cells using 488-nm excitation and 580–650-nm emission with ~25% total laser output for 5 s. 4. Scan field with reduced zoom factor. Pre-scanned area appears at least fivefold brighter.

3.6.2. Preparationdependent Labeling with MitoTracker® Red CM-H2 Xros (Fig. 6; See Note 18)

1. Cultivate cells in complete media to stationary growth phase (48 h). 2. Label cells for 10 min in suspension with 1 mg/ml MitoTracker® Red CM-H2XRos and mount on agar-coated slides, or 2b.Mount 1 ml on standard glass slides. 3. Image cells 5 min after preparation, using 543-nm excitation and 550–650-nm detection. 4. Depending on preparation conditions, fluorescence appears in mitochondria (agarose-coated slide) or spread throughout the cytosol (standard slide).

3.6.3. Laser-induced Morphological Alterations of FM4–64-Labeled Vacuoles (Fig. 7; See Note 20)

1. Cultivate cells to log phase in complete media (8 h). 2. Mount cells on slides coated with agar containing 1 mg/ml FM4–64. 3. Incubate for 60 min, to allow for incorporation of dye into vacuolar membranes. 4. Image cells at 488-nm excitation and 500-nm long-pass filter detection, with 25% max laser power; high-speed scanning and 8× averaging (1 s/frame). 5. Sequential recording of images displays extensive morphological alterations (fusion) of vacuoles, and appearance of rounded vacuoles with multiple inclusions.

88

Wolinski, Natter, and Kohlwein

Fig. 4. Visualizing protein and dye dynamics using bleaching technology. (a) Quantitative FRAP analysis of the influx of Nile Red molecules into yeast cells. Lipid droplets in stationary phase yeast cells were labeled with the lipophilic dye, Nile Red. Cells were mounted on standard microscope slides without washing, leaving some (nonfluorescent) dye in the surrounding media. Entire yeast cells were simultaneously illuminated with high laser intensity for 2 s (region of interests (ROIs) are labeled by white circles). The time-lapse record revealed a significant and uniform recovery of Nile Red fluorescence over time, indicating a continuous uptake of the dye into yeast cells. The cell in the middle of the image represents an unbleached control cell. (b) Qualitative FRAP analysis of Cox4-GFP labeled mitochondrial structures. Fluorescently labeled mitochondria in a bud were bleached with high laser intensity for 3 s (ROI indicated by white circle). 60 s after bleaching of mitochondrial Cox4-GFP in the bud, a significant recovery of fluorescence was detected in the bleached area, with a concomitant decrease of Cox4-GFP fluorescence in the mother cell. Images represent maximum-intensity projections of multiple optical sections acquired with high scan speed. Note the significant morphological alterations of mitochondrial structures of the unbleached control cells over time. (c) Qualitative FLIP analysis of Elo3-GFP labeled endoplasmic reticulum. A small part of the fluorescently labeled nuclear ER (ROI indicated by white circle) was continuously bleached for 10 s. The first postbleach image (1 s) shows a loss of fluorescence of the entire nuclear ER. In contrast, cortical ER structures remain unaffected by the bleaching process indicating the existence of a barrier for Elo3-GFP diffusion between nuclear and cortical ER. The gradual loss of fluorescence intensity of the non-bleached control cell is caused by photobleaching during measurement and has to be considered for a quantitative analysis of fluorescence recovery. Bar 5 mm.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy 89

90

Wolinski, Natter, and Kohlwein

Fig. 5. Laser-induced increase of DASPMI fluorescence. (a) Yeast cells were cultivated in complete media to stationary growth phase and labeled with 1 m/ml DASPMI. A field of view was imaged with increased zoom factor at 488 nm excitation; the laser source was set to ~20% of total output power. (b) Subsequent scan of the same field of view with decreased zoom factor. The scan revealed a significant increase of DASPMI fluorescence in cells of the previously scanned area. The rectangular mask indicates the scanned region shown in (a). Single optical sections; fluorescence and DIC transmission images. Bar 10 mm.

3.6.4. PreparationDependent Localization of Acc1-GFP (Fig. 8, See Note 21)

1. Cultivate cells expressing chromosomally integrated Acc1GFP to late log or to stationary phase. 2. (a) Mount log phase or stationary phase cells on standard microscope slides under cover slips or (b) Mount log phase cells on agar-coated slides. 3. Image GFP-distribution over time, using standard settings (488-nm excitation, 500–550-nm band-pass detection). Immediately after preparation, Acc1-GFP is present exclusively in the cytosol of log-phase cells imaged both on agar-coated slides and on standard microscope slides.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

91

Fig. 6. Labeling “artifacts” of a vital dye due to different preparative conditions. Yeast cells were cultivated to early stationary growth phase (12–24-h cultures) and labeled with MitoTracker® Red CM-H2XRos for 10 min. (a) After labeling, cells were mounted on agar-coated slides and imaged after 15 min. The single optical section shows fluorescence distribution typical for active mitochondria, at the cellular periphery. (b) After labeling, cells were mounted on a standard microscope slide without agar support and imaged after 5 min. In these cells, the vital dye brightly fluoresces in the cytosol and mitochondria are barely detectable. Bar 10 mm.

4. After 5 min, on standard slides, Acc1-GFP forms foci, which are present in more than 50% of the cells after 10 min; cells prepared on agarose-coated slides retain cytosolic Acc1-GFP localization over extended periods of time. (a) Inclusion of an air bubble in a cell preparation on standard slides shows that foci are absent about 50 mm distal to the air bubble, demonstrating the sensitivity and dependence of Acc1-GFP localization on the microenvironment, during microscopy.

92

Wolinski, Natter, and Kohlwein

Fig. 7. Illumination of labeled vacuoles induces rapid morphological changes. Log-phase cells were mounted on slides coated with agar containing 1 mg/ml FM4–64. After incubation for 30 min in the dark, cells were sequentially imaged using fast-scan mode and 8× averaging (about 2 s/frame ~180 ms dwell time in the displayed scanning areas; #1–#20 indicate the frame sequence). Morphological alterations of some vacuoles (marked by white arrows in the lower magnification panel) become evident after seven scans (total laser dwell time in that area 1.2 s). After 20 scans, vacuoles appear completely rounded in all cells of the scanned area. Bar 5 mm.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

93

Fig. 8. Changes of Acc1-GFP localization due to the microenvironment on the slide. (a) Yeast cells cultivated in complete media to late logarithmic growth phase were mounted on a standard microscope slide and imaged over time. Immediately after preparation, Acc1p-GFP shows a cytosolic distribution. Within minutes, Acc1-GFP localization changes and appears in bright punctae (foci) in most cells, when incubated on a standard slide. (b) Same cell suspension as in (a), but mounted on agar-coated slides (YPD). Aggregation of Acc1p-GFP is prevented under these microscopic conditions. (c) Relocalization of Acc1p-GFP due to an oxygen gradient within the preparation. Stationary phase cells are mounted on standard slides next to an air bubble. Cells in close vicinity to the air bubble (within ~50–60 mm) disassemble Acc1-GFP foci and the protein becomes dispersed in the cytosol, whereas cells in greater distance to the air bubble retain Acc1GFP foci, after 10 min of incubation. Maximum-intensity projections of multiple (9–13) optical sections. DIC transmission images on the right and fluorescence images middle and left. Bar 10 mm.

94

Wolinski, Natter, and Kohlwein

4. Notes 1. If only a few samples of cells are to be imaged, we recommend a protocol for cell preparation based on a method first described by Allan (15) and adapted for microscopy (1). This protocol involves isolation of virgin daughter cells from stationary phase cultures by density gradient centrifugation. This procedure results in a highly homogeneous population of cells that may enter vegetative growth in a synchronous way. 2. There is a strong bias in the visual inspection towards “bright” cells but serious consideration must be given as to the significance of this observation, in particular to whether the overall pattern of protein distribution in a cell depends on the level of fluorescence intensity. We find it re-assuring that localization patterns appear in most cases independent of the intensity (expression) level, demonstrating typically robust targeting. Also, plasmid loss under nonselective conditions is often over-emphasized; during the period of microscopic experiments, we find it very useful to cultivate plasmid-harboring cells in complete media (2–6 h), in order to stimulate growth and FP formation. Plasmid-loss is negligible during that period but the FP signal typically improves significantly. A strategy for introducing a chromosomally integrated FPtagged “query” sequence into the deletion strain collection based on “synthetic genetic array” technology (11, 13) is clearly the method of choice to screen mutant collections for alterations of specific localization patterns. 3. If the screen is designed to identify cells in the mutant collection with altered localization patterns or morphologies, one may consider fixing the cells prior to microscopy, in order to freeze cell morphologies at a similar stage of development, and thus allowing for comparison of cells in a similar state of growth. This trick is also of advantage for screening mutant populations with vital dyes since mild fixation typically eliminates differences in label uptake. However, imaging based screens are also suitable to identify alterations in organelle dynamics and inheritance, which obviously requires more time-consuming repetitive analyses of the same population of living cells. 4. In order to reduce the impact of light on the cells, illumination should be as short as possible and with lowest intensity possible; factors, which are mainly determined by the intensity of the fluorescence signal and the sensitivity of the microscope system. Confocal microscopy setups – which we prefer

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

95

for high-resolution yeast imaging – now provide increased scanning speed and improved (cooled) photodetectors which enable high-speed, high-resolution imaging with excellent signal-to-noise ratios and low electronic background. In our hands, confocal systems are advantageous due to their short dwell-times in a particular area of the specimen, thus reducing bleaching or physiological interference. The rather slow imaging speeds of conventional confocal systems (typically ~1 s/frame) compared to video/deconvolution systems (~10 ms/frame) are now challenged by the introduction of resonant scanner systems that provide video-rate imaging capabilities with typical confocal resolution. 5. Since live yeast cells are typically maintained in aqueous media during microscopy there is a significant loss in brightness (due to diffraction), and resolution. Water or glycerol immersion objectives with an adjustable collar to correct for variations in coverslip thickness (which is critical especially for water and glycerol immersion objectives) are preferable for imaging yeast cells in aqueous solutions. Even with a similar nominal numerical aperture for various objectives, resolution is significantly determined by the type of immersion fluid and demonstrates significant advantages of water or glycerol immersion lenses in comparison to typical (and cheaper) oil immersion lenses. 6. For publication, only minimal enhancements of images are considered appropriate, such as adjustments of brightness and contrast. Other types of image manipulation, such as background correction, nonlinear image adjustments etc. are considered inappropriate and not acceptable for publication, unless explicitly explained (see e.g., editorial policies, e.g., J. Cell Biology: http://www.jcb.org/misc/ifora. shtml#image_aquisition) (16). 7. The gene for nurseothricin resistance (ClonNAT) contains a GC rich region. When standard PCR conditions with proof reading enzymes fail, we use a PCR kit for GC rich regions (e.g., GC-rich PCR System, Roche Diagnostics, Germany) and a two-step PCR protocol with the following cycling conditions: 8×(95°C/30 s – 55°C/20 s – 72°C/3 min), 22×(95°C/30 s – 65°C/20 s – 72°C/3 min) in a GeneAmp 9600 thermocycler (Applied Biosystems). 8. To reduce potential problems resulting from (weak) GFP dimerization we are routinely using monomeric GFP (17) which we have introduced into vector pFa6a-GFP(S65T)kanMX4 (18). In this plasmid, the kanMX cassette was replaced by the ClonNAT cassette, resulting in the vector pKN082. Chromosomal integration of the GFP-ClonNAT PCR fragment is performed as described.

96

Wolinski, Natter, and Kohlwein

9. Growth conditions largely determine staining efficiency of vital dyes and GFP expression levels. Thus, great care has to be taken to maintain reproducible growth conditions. We have previously described a method for cell preparation for small scale imaging experiments with vital dyes, which is based on a centrifugation-based method for isolation of unbudded daughter cells from stationary phase cultures that are particularly suited for microscopic analysis (1). Precultivation of cell colonies for large-scale investigations, i.e., imaging-based mutant screenings using vital dyes or GFPlabeled cells (see Subheading 3.1), must by nature be less laborious and is, therefore, restricted to (robotics-based) replicating to appropriate media plates, with the disadvantage of representing more heterogeneous cell populations. 10. Make sure that colonies resulting from SGA protocol are not contaminated with unlabeled cells remaining from an incomplete elimination by the various selection steps. The repetitive replication process may help to obtain more homogeneous cell populations and eliminate dead cells. 11. Complete YPD is rich in fluorescent components, which may obscure weak GFP fluorescence signals. Thus, minimal media are preferable for fluorescence microscopy. 12. Long-term observation of cells requires adequate immobilization, which is readily achieved using the agar technique. The cells are typically arranged as a monolayer in a liquid film right underneath the cover slip. Bleaching experiments (FRAP) may require mounting of cells without agarose (agar) in order to avoid drifting of the focal plane (see also Note 15 below). Agarose-immobilization is required for long-term live cell imaging and for maintaining “physiological conditions”). 13. Images are typically stored using the Leica microscope software or a relational database. Image processing and evaluation follows in-house protocols and software tools, i.e., at least two independent viewers and automated evaluation of fluorescence intensity distribution. Very recently, first applications of automated image analysis and feature extraction have been published (2). 14. Typically, objectives with 63× or 100× magnification and high numerical apertures (1.2–1.4) are to be used for yeast imaging. Despite comparable (or perhaps even somewhat lower) numerical apertures, glycerol and water immersion optics with adjustable collar ring may yield better resolution than oil immersion lenses for live yeast cell imaging. This is due to the aqueous environment and the more glycerolcompatible diffraction index of the cells. The impact of nonrefractive index-matched optics becomes noticeable already

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy

97

a few mm below the cover slip, about half way through the yeast cells. Spectral fluorescence detection allows for optimized adjustment of the detection wavelength range and efficient suppression of background fluorescence. 15. Cell preparation typically poses some stress on the cells that may result in altered labeling. Also, the agar layer may shrink somewhat due to water evaporation, which results in cell displacement. 16. Although photobleaching is a widely used approach to address protein and organelle dynamics several reports have been published pointing out limitations, pitfalls, and potential artifacts generated by this technology that need to be carefully considered (19–21). 17. Microscopists are aware of multiple experimental problems during live cell imaging. Most prominently, fluorescence may bleach due to excess light absorption by the dye and subsequent chemical modification, resulting in a loss of the ability to emit photons. Other phenomena like fluorescence saturation, a process that limits fluorescence emission, is frequently observed as well. As a consequence, fluorescence emission may be much higher at lower excitation intensity, stressing that “more light” does not necessarily result in more fluorescence. Obviously, numerous types of molecules in an unstained cell may absorb photons as well, which may significantly interfere with a cell’s physiology and ultimately result in slower growth and potentially in cell death. In this respect, UV light is much more potent and cytotoxic, and dyes absorbing in the red or far-red spectrum are, therefore, becoming more and more popular. The application of twophoton microscopy, which potentially circumvents UV toxicity by using high-intensity pulsed IR lasers that may also excite UV-dyes, however, is obviously limited (in yeast) due to excess heat generation. As a general rule, use of fluorescence dyes typically renders cells much more sensitive to light, thus dictates to keep excitation light intensity, illumination time and concentration of the label as low as possible. 18. We have frequently observed that over time (i.e., by timelapse imaging) fluorescence intensity of a specimen may increase, rather than decrease (bleaching). Typically, such phenomena are not readily observed in epifluorescence, which relies on simultaneous illumination of an entire field of view over extended periods of time (seconds). Fluorescence activation may be due to light-induced stimulation of dye uptake (i.e., by interfering with the cell’s physiology), or alterations of dye concentration and interaction with other molecules within a particular structure.

98

Wolinski, Natter, and Kohlwein

19. The microenvironment during imaging is of critical importance for obtaining reliable imaging information. Especially, mitochondria or other oxygen-dependent cellular entities may react very sensitively to induced anaerobiosis in standard preparations. 20. Vacuoles are a very dynamic membrane-bound organelle, undergoing rapid fusion and fission processes. FM4–64 has been used extensively to monitor endocytosis (12) and vacuolar dynamics. 21. Cells prepared for microscopy are subject to microenvironmental conditions that may affect organelle morphology and protein localization. Thus, cell preparation and maintenance under the microscope need careful consideration in all live cell imaging experiments.

Acknowledgements Work in our laboratory is supported by grants from the Austrian Federal Ministry for Science and Research (GEN-AU program, projects GOLD – Genomics of Lipid-associated Disorders, and FLUPPY – Fluorescence-based analysis of protein–protein interactions in yeast), and the Austrian Science Fund, FWF (project SFB Lipotox F3005). References 1. Wolinski H., Kohlwein S.D. (2008). Membrane Trafficking, in Methods in Molecular Biology Vol 457, A. Vancura, ed. Humana Press, pp. 151–163. 2. Chen S.C., Zhao T., Gordon G.J., Murphy R.F. (2007). Automated image analysis of protein localization in budding yeast. Bioinformatics (Oxford, England) 23(13), 66–71. 3. Carpenter A.E., Jones T.R., Lamprecht M.R., et al. (2006). CellProfiler: image analysis software for identifying and quantifying cell phenotypes. Genome Biology 7(10), R100. 4. Lamprecht M.R., Sabatini D.M., Carpenter A.E. (2007). CellProfiler: free, versatile software for automated biological image analysis. BioTechniques 42(1), 71–5. 5. Huh W.K., Falvo J.V., Gerke L.C., et al. (2003). Global analysis of protein localization in budding yeast. Nature 425(6959), 686–91. 6 . Kals M ., Natter K ., Thallinger G.G ., Trajanoski Z., Kohlwein S.D. (2005). YPL.db2: the yeast protein localization database, version 2.0. Yeast 22(3), 213–8.

7. Natter K., Leitner P., Faschinger A., Wolinski H., McCraith S., Fields S., Kohlwein S.D., et al. (2005). The spatial organization of lipid synthesis in the yeast Saccharomyces cerevisiae derived from large-scale green fluorescent protein tagging and high resolution microscopy. Molecular and Cellular Proteomics 4(5), 662–72. 8. Wiwatwattana N., Landau C.M., Cope G.J., Harp G.A., Kumar A. (2007). Organelle DB: an updated resource of eukaryotic protein localization and function. Nucleic Acids Research 35(Database issue), D810–4. 9. Kumar A., Cheung K.H., Tosches N., et al. (2002). The TRIPLES database: a community resource for yeast molecular biology. Nucleic Acids Research 30(1), 73–5. 10. Saito T.L., Ohtani M., Sawai H., et al. (2004). SCMD: Saccharomyces cerevisiae morphological database. Nucleic Acids Research 32(Database issue), D319–22. 11. Tong A.H., Boone C. (2006). Synthetic genetic array analysis in Saccharomyces cerevisiae. Methods in Molecular Biology 313, 171–92.

The Fidgety Yeast: Focus on High-Resolution Live Yeast Cell Microscopy 12. Vida T.A., Emr S.D. (1995) A new vital stain for visualizing vacuolar membrane dynamics and endocytosis in yeast. Journal of Cell Biology 128(5), 779–92. 13. Tong A.H., Evangelista M., Parsons A.B., et al. (2001). Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294(5550), 2364–8. 14. Boone C., Bussey H., Andrews B.J. (2007). Exploring genetic interactions and networks with yeast. Nature Reviews 8(6), 437–49. 15. Allen C., Buttner S., Aragon A.D., et al. (2006). Isolation of quiescent and nonquiescent cells from yeast stationary-phase cultures. Journal of Cell Biology 174(1), 89–100. 16. Rossner M., Yamada K.M. (2004). What’s in a picture? The temptation of image manipulation. Journal of Cell Biology 166(1), 11–5. 17. Zacharias D.A., Violin J.D., Newton A.C., Tsien R.Y. (2002) Partitioning of lipid-modified

18.

19.

20.

21.

99

monomeric GFPs into membrane microdomains of live cells. Science 296(5569), 913–6. Wach A., Brachat A., Alberti-Segui C., Rebischung C., Philippsen P. (1997). Heterologous HIS3 marker and GFP reporter modules for PCRtargeting in Saccharomyces cerevisiae. Yeast 13(11), 1065–75. Lippincott-Schwartz J., Altan-Bonnet N., Patterson G.H. (2003). Photobleaching and photoactivation: following protein dynamics in living cells. Nature Cell Biology, Suppl:S7–14. Pucadyil T.J., Chattopadhyay A. (2006). Confocal fluorescence recovery after photobleaching of green fluorescent protein in solution. Journal of Fluorescence 16(1), 87–94. Sbalzarini I.F., Mezzacasa A., Helenius A., Koumoutsakos P. (2005). Effects of organelle shape on fluorescence recovery after photobleaching. Biophysical Journal 89(3), 1482–92.

Chapter 6 A Genomic Approach to Yeast Chronological Aging Christopher R. Burtner, Christopher J. Murakami, and Matt Kaeberlein Summary Yeast is a useful model organism to study the genetic and biochemical mechanisms of aging. Genomic studies of aging in yeast have been limited, however, by traditional methodologies that require a large investment of labor and resources. In this chapter, we describe a newly-developed method for quantitatively measuring the chronological life span of each strain contained in the yeast ORF deletion collection. Our approach involves determining population survival by monitoring outgrowth kinetics using a Bioscreen C MBR shaker/incubator/plate reader. This method has accuracy comparable to traditional assays, while allowing for higher throughput and decreased variability in measurement. Key words: Longevity, Aging, Chronological life span, Yeast, Bioscreen, Stationary phase

1. Introduction Along with the nematode Caenorhabditis elegans and the fruit fly Drosophila melanogaster, the budding yeast has emerged as one of the primary invertebrate models used by scientists studying the molecular biology of aging (1). Two aging paradigms have been developed in yeast: replicative and chronological (Fig. 1). Replicative aging refers to the mitotic capacity of a yeast mother cell and is defined as the number of daughter cells produced by a mother cell prior to senescence (2). Chronological aging describes the survival of yeast cells in a nondividing, quiescentlike state and is generally defined by the length of time a cell can survive in stationary phase (3). The ability to specifically monitor the effect of genetic and environmental perturbations

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_6, © Humana Press, a part of Springer Science + Business Media, LLC 2009

101

102

Burtner, Murakami, and Kaeberlein

Fig. 1. Schematic for Yeast Replicative and Chronological Aging. (a) Replicative life span in yeast is measured by the number of mitotic divisions that can arise from a single mother cell. Replicative viability is calculated as the mean number of daughters produced from mothers of a particular strain background before senescence. (b) Chronological life span as measured by the length of time cells in a stationary culture can remain viable. Viability is calculated by the fraction of the culture able to reenter the cell cycle after an extended state of quiescence. Figure modified from (16).

on the longevity of both dividing and nondividing cell types is a major advantage of yeast as a model for studies of aging. Several dozens of genetic and environmental factors have been identified which influence longevity in yeast, at least some of which play a similar role in multicellular eukaryotes (4). For example, dietary restriction, which has long been known to increase life span in multicellular eukaryotes, such as worms, flies, and rodents, also increases both replicative life span (5–7) and chronological life span (8, 9) in yeast. Similarly, decreased TOR signaling (a response to dietary restriction) is also known to increase both chronological (10) and replicative life span (11) in yeast, as well as life span in worms and flies. Additional genes that play an apparently conserved role in modulating longevity in yeast and other organisms include SIR2(12), SCH9(13–15), and protein kinase A (7, 15). The fact that multiple interventions promoting longevity in yeast also promote longevity in multicellular organisms strongly suggests that yeast can be a useful model for at least some aspects of aging in higher eukaryotes (16). Although studies of aging in yeast have been extremely informative regarding genetic factors that influence longevity, these studies have been limited by the relatively tedious and time consuming nature of the traditional assays for measuring replicative and chronological life span. Higher-throughput approaches have

A Genomic Approach to Yeast Chronological Aging

103

recently been developed for life span determination, however, and are being applied toward genomic screens for increased longevity. An ongoing screen of the yeast ORF deletion collection utilizes an iterative method for identifying single-gene deletion mutants that have increased replicative life span (11). More recently, a qualitative high-throughput method was described for assaying chronological life span in 96-well plates (10). Both of these methods have been described in detail elsewhere (1, 4, 10, 11, 17), and we refer the interested reader to these references. This chapter will describe a newly developed approach for measuring chronological life span that greatly improves upon the quantitative power of prior methods, while still providing relatively high-throughput capacity. This method utilizes the kinetics of cell growth exhibited by a subset of the aging culture to monitor viability of the population over time (e.g., chronological aging in stationary phase). The validity of such an approach relies on the assumption that, given fixed environmental growth conditions, doubling time, and initial growth state, the relative OD of two (or more) different yeast cultures at any point during subsequent outgrowth will be determined by the relative number of viable cells present at the initial time point. The fraction of viable cells in a particular sample relative to the control sample can be determined based on the corresponding growth curves by the formula: vn =

1 ( Δt n / d )

2

,

where v is the viability in sample n relative to control, Δtn equals the time shift between the outgrowth curves of sample n and the control sample (Fig. 2), and d equals the doubling time of the

Fig. 2. Determining relative viability from outgrowth curves. The relative number of viable cells in two samples can be determined from the kinetics of outgrowth based on the given formula, where vn is the relative viability of the two samples, d is the doubling time of the strain (assumes samples A and B have identical doubling times), and Δtn is the time shift between the two outgrowth curves.

104

Burtner, Murakami, and Kaeberlein

strain (determined by the maximal slope of the semi-log plot of OD as a function of time). The primary advance associated with this new protocol is the use of a Bioscreen C MBR (Growth Curves, USA) machine to monitor outgrowth based on optical density (OD). The Bioscreen C MBR machine is an automated shaker/incubator/plate reader useful for determining growth kinetics of yeast strains under a variety of conditions (18, 19). This method has been demonstrated to be accurate by comparison with the traditional colony forming unit assay (CFU), with an increase in precision as determined by lower variance in measurement (20). We are currently using the method described in the following sections to quantitatively determine the chronological aging properties for each strain in the yeast ORF deletion collection (21). This protocol could be easily adapted for other studies of yeast aging or studies examining the sensitivity of yeast cells to a variety of chemical or environmental stressors.

2. Materials 1. Preparation of YPD agar plates (500 mL): In a 1-L Erlenmeyer flask combine 5 g yeast extract (BD, Bacto™ Yeast Extract) and 10 g peptone (BD, Bacto™ Peptone) in 240 mL diH2O and add a stir bar. In another 1-L Erlenmeyer flask, combine 10 g agar with 235 mL diH2O. Cover both flasks with aluminum foil, and autoclave for 45 min. Allow the yeast extract/ peptone (YEP) solution to cool on a stir plate with gentle agitation, and slowly add the agar solution to the YEP when the temperature of the agar is warm to the touch (~55°C). Add 40 mL of filter sterilized 50% glucose to the medium (2% final concentration), and dispense approximately 25 mL into individual 100-mm Petri dishes. 2. Preparation of liquid YPD medium (500 mL): In a 2-L flask, combine 450 mL of diH2O, 5 g yeast extract (BD, Bacto™ Yeast Extract) and 10 g peptone (BD, Bacto™ Peptone), and autoclave for 45 min with a loose cap. Add 20 mL of 50% filter sterilized glucose (2% final concentration), and allow to completely cool to room temperature before use. 3. Preparation of liquid Synthetic Defined (SD) medium: Basic medium (B) is composed of 1.7 g/L Yeast Nitrogen Base -AA/-AS (BD, Difco™ Yeast Nitrogen Base) and 5.0 g/L

A Genomic Approach to Yeast Chronological Aging

105

Table 1 Preparation of amino acid powder mix and 10× stock for SD medium Component

Powder mix (g)

10× Stock (g/L)

Adenine

2.5

0.4

L-arginine

1.2

0.2

L-aspartate

6

1

L-glutamate

6

1

L-histidine

6

0.2

L-leucine

18

0.6

L-lysine

1.8

0.3

L-methionine

1.2

0.2

L-phenylalanine

3

0.5

L-serine

22.5

3.75

L-threonine

12

2

L-tryptophan

2.4

0.4

L-tyrosine

1.8

0.3

L-valine

9

1.5

Uracil

6

0.2

Add the gram quantities shown for each component and mix well. Make 10× stock by dissolving 16.55 g of powder mix per liter of Basic medium

(NH4)2SO4. The medium is then autoclaved and can be stored at room temperature. The amino acid powder mix is made by combining the individual amino acids (Sigma, St. Louis, MO) in gram amounts listed in the first column of Table 1. A 10× amino acid stock is made by adding 4.14 g of the powder mix to a final volume of 250 mL B, and filter sterilized. The 10× amino acid stock should be stored at 4°C and kept from exposure to light. To make 500 mL of liquid SD medium, add 50 mL of 10× amino acid stock and 20 mL of filter sterilized 50% glucose to 430 mL of B. The final concentration of each component in the SD medium is provided in Table 2.

106

Burtner, Murakami, and Kaeberlein

Table 2 Final composition of synthetic defined medium used for yeast chronological life span analysis of the deletion collection Carbon source Glucose

Base pre-mix 20 g/L

Amino acids

Potassium phosphate

1 g/L

Magnesium sulfate

500 mg/L

Sodium chloride

100 mg/L

L-arginine

20 mg/L

Calcium chloride

100 mg/L

L-aspartate

100 mg/L

Biotin

0.002 mg/L

L-glutamate

100 mg/L

Pantothenate

0.4 mg/L

L-histidine

20 mg/L

Folate

0.002 mg/L

L-leucine

60 mg/L

Inositol

2 mg/L

L-lysine

30 mg/L

Niacin

0.4 mg/L

L-methionine

20 mg/L

PABA

0.2 mg/L

L-phenylalanine

50 mg/L

Pyridoxine, HCl

0.4 mg/L

L-serine

375 mg/L

Riboflavin

0.2 mg/L

L-threonine

200 mg/L

Thiamine, HCl

0.4 mg/L

L-tryptophan

40 mg/L

Riboflavin

0.2 mg/L

L-tyrosine

30 mg/L

Thiamine, HCl

0.4 mg/L

L-valine

150 mg/L

Boric acid

0.5 mg/L

Copper sulfate

0.04 mg/L

Potassium iodide

0.1 mg/L

Other components Adenine

40 mg/L

Ferric chloride

0.2 mg/L

Uracil

20 mg/L

Manganese sulfate

0.4 mg/L

Ammonium sulfate

5 g/L

Sodium molybdate

0.2 mg/L

Zinc sulfate

0.4 mg/L

3. Methods 3.1. Preparation of Strains for Life Span Analysis

The Bioscreen C MBR machine can accommodate two 100-well Honeycomb plates simultaneously (Fig. 3). In our experience, we have found that chronological life span determination of triplicate

A Genomic Approach to Yeast Chronological Aging

107

Fig. 3. Honeycomb plates. The Bioscreen C MBR machine uses 100-well Honeycomb plates designed to provide uniform heating in each well. Two plates can be accommodated per run, allowing for a maximum throughput of 200 assays per day per machine.

biological replicates is sufficient to discriminate between relatively small changes in survival. Therefore, we determine the chronological life span of 64 deletion strains per experiment, allowing for triplicate wild type controls to be included in each Honeycomb plate. Based on differential survival in 96-well plates versus culture tubes, we have chosen to age the strains in culture tubes on a rotating drum. Alternative culture conditions could easily be adapted. 1. Cells from 64 deletion strains are obtained from the Yeast ORF Deletion Collection and patched onto a YPD plate. The patch is streaked for single colonies, and allowed to grow at 30°C for 48 h. 2. A single colony from each strain is inoculated into a culture tube containing 5 mL liquid YPD. The liquid cultures are grown at 30°C overnight in a roller drum. 3. To generate the aging cultures for the Bioscreen experiments, 50 μL of the YPD overnight culture is inoculated into 5 mL of SD medium in triplicate (1:100 dilution). Aging cultures are kept in the rotating roller drum at 30°C for the duration of the experiment.

108

Burtner, Murakami, and Kaeberlein

3.2. Obtaining Bioscreen Outgrowth Curves at Each Age Point

To measure the loss in viability over time, an aliquot of the aging culture is periodically challenged for its ability to re-enter the cell cycle in the presence of a rich medium. This is accomplished by inoculating a small volume from the aging culture into fresh medium and monitoring outgrowth using a Bioscreen C MBR machine. The software interface included with the Bioscreen C MBR, called Easy Bioscreen Experiment, allows the user to define the temperature of the incubation, the interval at which OD measurements are taken, and the degree and duration of agitation (shaking) of the Bioscreen plate, to keep the culture aerated and the cells in suspension. The design of the Honeycomb plate is optimized to keep heat equally distributed across the plate, so all wells are incubated similarly. The OD readings are given as an output in a tab-delimited format, which can be plotted to visualize outgrowth curves for every aging yeast strain (with a maximum of 200 strains per experiment). 1. In the Easy Bioscreen Experiment window, define the settings for the incubation: a) Indicate the number of samples to be measured (200). b) In the Filter pulldown menu, select Wideband (420– 580 nm). c) Set the temperature to 30°C. d) Select the Settings menu, and choose Shaking. Set the shaking amplitude to high in the pulldown menu, and check the box below for continuous shaking. e) Set measurement interval to 30 min. f) Set experiment length to 24 h. 2. Prepare the Bioscreen Honeycomb plates by filling each well with 145 μL liquid YPD (see Note 3). A standard multichannel pipettor will not align perfectly with the spacing of the wells in the Honeycomb plate, but we find that we can use five tips at a time to load the plate. Alternatively, an expandable multichannel pipettor can be used. 3. Remove aging cultures from the roller drums within the 30°C incubators. 4. Just prior to inoculation, vortex each culture briefly to ensure all cells are evenly suspended. Inoculate 5 μL of each culture into one well of the Honeycomb plate using appropriate sterile technique. Note: For consecutive age-points within the same experiment, maintain a consistent order of inoculation from aging cultures into the Honeycomb plate. In other words, the same aging culture should always be loaded in the same well # at each age-point. This greatly simplifies subsequent data analysis, as described below. 5. After the plate is loaded, place the Honeycomb plate into the Bioscreen, with well #1 in the upper-left corner of the plate

A Genomic Approach to Yeast Chronological Aging

109

holder. Begin the run by clicking Start in the Easy Bioscreen Experiment window. The Bioscreen machine will progress through a series of calibration steps before taking the first absorbance reading (see Notes 1 and 2). 6. Repeat these steps at day 2, 4, 6, 9, 11, and 13 of the experiment. 3.3. Quantifying Doubling Time and Normalizing Data for Each Age-Point

The Bioscreen C MBR machine gives the output of the OD data as a tab delimited text file that is compatible with Microsoft Excel and a variety of other software. The first column of the output data corresponds to the time at which each OD reading was obtained. Subsequent columns correspond to the OD measurement obtained from each well of the Honeycomb plates at each time point. We use a MATLAB script to analyze the raw data as described below and plot graphs in Microsoft Excel. 1. Obtain the data from the Bioscreen C MBR machine as a tabdelimited text file. 2. Subtract the OD corresponding to YPD without inoculated cells from each measured OD value. The OD of YPD can be obtained from a well containing YPD without inoculated cells (generally ~0.14) and may vary slightly from experiment to experiment. 3. Calculate doubling time (d) for each column from the maximal slope of the equation defined by the natural logarithm of OD as a function of time. In practice, for each column, we calculate a linear equation of ln(OD) versus time for every two consecutive time points and use the median of the five lowest doubling time values as an estimate of growth rate. Since the length of time between each measurement is constant, the doubling time corresponding to any two consecutive OD measurements can be calculated as, ln(2) , ⎛ ln(OD2 ) − ln(OD1 ) ⎞ ⎜⎝ ⎟⎠ t 2 − t1 where OD1 and OD2 are any two consecutive OD readings, and t2 and t1 are two consecutive time points (t2 − t1 = 30 min, in this case). For most strains, the value obtained will be a doubling time of 85–95 min. 4. After doubling time has been calculated for each column, normalize the data for each column by removing the first entry in each column and subtracting the value of the second entry from every entry in the column. This controls for potential differences in initial cell density or optical density variation among the wells.

110

Burtner, Murakami, and Kaeberlein

3.4. Determining Chronological Life Span from Bioscreen Outgrowth Data

After the raw data for each age-point in the experiment has been treated as described in Subheading 3.4, a survival plot can be generated for each well. This requires merging the data files from individual age-points in a manner such that each well (corresponding to one biological replicate of one yeast strain) can be analyzed independently over multiple age-points. For simplicity, the following steps assume that the data for a single sample (which will correspond to a single well # if the Honeycomb plate was loaded properly at each age-point) has been organized in a spreadsheet such that the first column contains the time values, the second column contains the initial age-point data (e.g., day 2), the third column contains the second age-point (e.g., day 4), and so forth. 1. Plot the data. This will allow visualization of the shift in the outgrowth curve as a function of age (Fig. 4a, b) (see Note 5). 2. For each age-point, calculate the time shift (Δt) in the outgrowth curve relative to the initial age point. In our experience, there is relatively little variation in Δt for any two outgrowth curves between OD values of 0.1 and 0.5. Calculation of Δt can be accomplished by choosing one or more fixed OD values, estimating the time at which each growth curve reached each OD, and determining the difference in those time values. The time that a particular growth curve achieves a fixed OD value can be calculated from the linear regression equation corresponding to ln(OD) as a function of time between the two time-points bracketing the fixed OD value. 3. From these data a mortality curve can be generated. Define the first age-point as 100% survival. For each subsequent agepoint, calculate the percent survival based on the time shift Δtn, by the formula sn =

1 ( Δt n / dn )

2

,

where sn is the fractional survival at age-point n, Δtn equals the time shift between the growth curve at age-point n and the growth curve at the initial age-point (e.g., day 2), and dn equals the doubling time of the strain (Fig. 4c) (see Note 6). 3.5. Statistical Analysis of Changes in Chronological Survival

1. When performing a screen for chronological survival it is desirable to quantify the difference in life span between two strains. There are two general approaches for accomplishing this. One could measure the median and maximum life spans, typically defined as the age at which 50% of the population is no longer viable and the median of the upper decile of survival (~95% senescence), respectively. However, these two measurements can vary independently of each other, and, in our experience, maximum life span measurement can be difficult to determine

Fig. 4. Representative chronological life span outgrowth and survival curves. Outgrowth curves generated by the Bioscreen C MBR and normalized according to Subheading 3.4 for the (a) WT BY4743 diploid and (b) an isogenic homozygous deletion strain from the ORF Deletion Collection (vma2D). (c) Mortality curves for each of the strains were created according to Subheading 3.5. Deletion of Vma2, a subunit of vacuolar ATPase, results in a shorter life span compared to the wild type. (d) Survival integral for BY4743 and vma2D. Error bars are standard deviation of three biological replicates.

A Genomic Approach to Yeast Chronological Aging 111

112

Burtner, Murakami, and Kaeberlein

in some chronological aging experiments. Thus, we utilize an alternative approach to estimate survival in which the integral of the survival curves (survival integral, SI) are compared (10). 2. Once survival data has been generated for each well over the entire course of the experiment, the area under the survival curve (SI) can be estimated by the formula: n ⎛s +s ⎞ SI = ∑ ⎜ n −1 n ⎟ (agen − agen −1 ), 2 ⎠ 2 ⎝

where agen is the age-point (e.g., refs. 2, 4, 6, 9, 11, and 13) and sn is the survival value at that age-point (see Note 4). For instance, if viability is 100% on day 2 and 75% on day 4, then the area under the survival curve between those two age-points is 1.75. The sum of this calculation for each pair of consecutive age-points in the experiment is the SI (Fig. 4d). 3. Once the SI has been determined for each sample, it is straightforward to calculate statistical values such as mean, median, and variance for each set of triplicate biological replicates. A t-test or similar analysis can be used for pairwise comparison of individual deletion strains against wild type (or other deletion strains), in order to determine whether there is a significant difference in SI.

4. Notes 1. We use YPD medium for outgrowth in the Bioscreen C MBR machine. This is done because growth rate is faster in YPD, allowing for a greater sensitivity to detect shifts in the growth curves. In addition, we have observed that outgrowth curves generated from cells grown in SD are more variable. 2. Occasionally the Bioscreen C MBR machine will display an F1.b2 error message indicating insufficient light from the bulb. This error indicates that the bulb may need to be replaced. It is a good idea to keep several spare bulbs on hand. 3. Although well-to-well variation is very small, over time, a fine plastic dust is produced by the shaking of the Bioscreen Honeycomb well plates. We have found that this particulate matter can obscure the optics, and produce noisy or inaccurate OD readings during the incubation. We therefore recommend cleaning the tray with ethanol and kimwipes regularly to prevent this problem. 4. Periodically during a chronological aging experiment, a subset of the cells will re-enter the cell cycle in the aging culture,

A Genomic Approach to Yeast Chronological Aging

113

a phenomenon referred to as “gasping” (22). This will be observed in the Bioscreen method as a leftward shift in the outgrowth curve at one or more later age-points, corresponding to an increase in the number of viable cells in the culture over time. Cultures where gasping has occurred should be removed from the analysis, unless viability is already low enough that subsequent age-points will not significantly influence SI. 5. Validation of chronological life span data obtained from the Bioscreen method can be periodically performed by plating for colony forming units from aging cultures and monitoring the change in colony formation. To date, we have not observed any significant differences between quantifying viability by colony forming units versus the Bioscreen method. 6. If relative chronological life span is determined by comparing the integral of the survival curves, it is important to maintain a fixed protocol of age-points. For our initial screen of the deletion collection, we monitor survival at day 2, 4, 6, 9, 11, and 13. The age-points used are somewhat arbitrary, but consistency across experiments is essential for fair comparison. These age-points work well for analysis of the ORF deletion collection. In some cases, (e.g., growth in 0.05% glucose), survival is still close to 100% after 14 days, so additional or alternative age-points may be desirable.

Acknowledgments The development of the method described here was supported by a pilot project grant to MK from the University of Washington Nathan Shock Center for Excellence in the Basic Biology of Aging Grant 5P30 AG013280. CRB is supported by National Institutes of Health Training Grant 5P30 AG013280. References 1. Kaeberlein, M. (2006). In Handbook of models for human aging. (Conn, P. M., Ed.), Elsevier Press, Boston, pp. 109–120. 2. Mortimer, R. K., and Johnston, J. R. (1959). Life span of individual yeast cells. Nature. 183, 1751–1752. 3. Fabrizio, P., and Longo, V. D. (2003). The chronological life span of Saccharomyces cerevisiae. Aging Cell. 2, 73–81. 4. Kaeberlein, M. (2006). Genome-wide approaches to understanding human ageing. Hum Genomics. 2, 422–428.

5. Jiang, J. C., Jaruga, E., Repnevskaya, M. V., and Jazwinski, S. M. (2000). An intervention resembling caloric restriction prolongs life span and retards aging in yeast. Faseb J. 14, 2135–2137. 6. Kaeberlein, M., Kirkland, K. T., Fields, S., and Kennedy, B. K. (2004). Sir2-independent life span extension by calorie restriction in yeast. PLoS Biol. 2, E296. 7. Lin, S. J., Defossez, P. A., and Guarente, L. (2000). Requirement of NAD and SIR2 for lifespan extension by calorie restriction in Saccharomyces cerevisiae. Science. 289, 2126–2128.

114

Burtner, Murakami, and Kaeberlein

8. Fabrizio, P., Gattazzo, C., Battistella, L., Wei, M., Cheng, C., McGrew, K., and Longo, V. D. (2005). Sir2 blocks extreme life-span extension. Cell. 123, 655–667. 9. Smith Jr, D. L., McClure, J. M., Matecic, M., and Smith, J. S. (2007). Calorie restriction extends the chronological lifespan of Saccharomyces cerevisiae independently of the Sirtuins. Aging Cell. 6, 649–62. 10. Powers, R. W., 3rd, Kaeberlein, M., Caldwell, S. D., Kennedy, B. K., and Fields, S. (2006). Extension of chronological life span in yeast by decreased TOR pathway signaling. Genes Dev. 20, 174–184. 11. Kaeberlein, M., Powers, R. W., 3rd, Steffen, K. K., Westman, E. A., Hu, D., Dang, N., Kerr, E. O., Kirkland, K. T., Fields, S., and Kennedy, B. K. (2005). Regulation of yeast replicative life span by TOR and Sch9 in response to nutrients. Science. 310, 1193–1196. 12. Kaeberlein, M., McVey, M., and Guarente, L. (1999). The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes Dev. 13, 2570–2580. 13. Fabrizio, P., Liou, L. L., Moy, V. N., Diaspro, A., SelverstoneValentine, J., Gralla, E. B., and Longo, V. D. (2003). SOD2 functions downstream of Sch9 to extend longevity in yeast. Genetics. 163, 35–46. 14. Fabrizio, P., Pletcher, S. D., Minois, N., Vaupel, J. W., and Longo, V. D. (2004). Chronological aging-independent replicative life span regulation by Msn2/Msn4 and Sod2 in Saccharomyces cerevisiae. FEBS Lett. 557, 136–142. 15. Fabrizio, P., Pozza, F., Pletcher, S. D., Gendron, C. M., and Longo, V. D. (2001). Regulation of longevity and stress resistance by Sch9 in yeast. Science. 292, 288–290. 16. Kaeberlein, M., Burtner, C. R., and Kennedy, B. K. (2007). Recent developments in yeast aging. PLoS Genet. 3, e84.

17. Kaeberlein, M., and Kennedy, B. K. (2005). Large-scale identification in yeast of conserved ageing genes. Mech Ageing Dev. 126, 17–21. 18. Malathi, K., Higaki, K., Tinkelenberg, A. H., Balderes, D. A., Almanzar-Paramio, D., Wilcox, L. J., Erdeniz, N., Redican, F., Padamsee, M., Liu, Y., Khan, S., Alcantara, F., Carstea, E. D., Morris, J. A., and Sturley, S. L. (2004). Mutagenesis of the putative sterol-sensing domain of yeast Niemann Pick C-related protein reveals a primordial role in subcellular sphingolipid distribution. J Cell Biol. 164, 547–556. 19. Tsuchiya, M., Dang, N., Kerr, E. O., Hu, D., Steffen, K. K., Oakes, J. A., Kennedy, B. K., and Kaeberlein, M. (2006). Sirtuin-independent effects of nicotinamide on lifespan extension from calorie restriction in yeast. Aging Cell. 5, 505–514. 20. Murakami, C. J., Burtner, C. R., Kennedy, B. K., and Kaeberlein, M. (2008). A method for high-throughput quantitative analysis of yeast chronological life span. J Gerontol A Biol Sci Med Sci. 63, 113–21. 21. Winzeler, E. A., Shoemaker, D. D., Astromoff, A., Liang, H., Anderson, K., Andre, B., Bangham, R., Benito, R., Boeke, J. D., Bussey, H., Chu, A. M., Connelly, C., Davis, K., Dietrich, F., Dow, S. W., El Bakkoury, M., Foury, F., Friend, S. H., Gentalen, E., Giaever, G., Hegemann, J. H., Jones, T., Laub, M., Liao, H., Davis, R. W., and et al. (1999). Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science. 285, 901–906. 22. Fabrizio, P., Battistella, L., Vardavas, R., Gattazzo, C., Liou, L. L., Diaspro, A., Dossen, J. W., Gralla, E. B., and Longo, V. D. (2004). Superoxide is a mediator of an altruistic aging program in Saccharomyces cerevisiae. J Cell Biol. 166, 1055–1067.

Chapter 7 Chemogenomic Approaches to Elucidation of Gene Function and Genetic Pathways Sarah E. Pierce, Ronald W. Davis, Corey Nislow, and Guri Giaever Summary The ~6,000 strains in the yeast deletion collection can be studied in a single culture by using a microarray to detect the 20 bp DNA “barcodes” or “tags” contained in each strain. Barcode intensities measured by microarray are compared across time-points or across conditions to analyze the relative fitness of each strain. The development of this pooled fitness assay has greatly facilitated the functional annotation of the yeast genome by making genome-wide gene-deletion studies faster and easier, and has led to the development of high throughput methods for studying drug action in yeast. Pooled screens can be used for identifying gene functions, measuring the functional relatedness of gene pairs to group genes into pathways, identifying drug targets, and determining a drug’s mechanism of action. This process involves five main steps: preparing aliquots of pooled cells, pooled growth, isolation of genomic DNA and PCR amplification of the barcodes, array hybridization, and data analysis. In addition to yeast fitness applications, the general method of studying pooled samples with barcode arrays can also be adapted for use with other types of samples, such as mutant collections in other organisms, siRNA vectors, and molecular inversion probes. Key words: S. cerevisiae, Drug-target identification, Functional assays, Chemogenomics, Gene networks, Genomics, DNA barcodes

1. Introduction The yeast deletion collection is a set of single-gene deletion strains in S. cerevisiae that covers almost every open reading frame (ORF) in the genome (96%) (1). The deletion collection was created to provide a pre-constructed, saturated set of gene-deletion mutants that can be used in the place of random-mutant libraries and individually constructed strains. In addition to fulfilling this

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_7, © Humana Press, a part of Springer Science + Business Media, LLC 2009

115

116

Pierce et al.

goal, the collection has also enabled the development of methods for studying all ~6,000 deletion strains in a single culture(1–3). These methods are made possible by the inclusion of unique 20 bp DNA “barcodes” or “tags” in each strain (Fig. 1). that enables relative strain concentrations to be compared by hybridizing pooled samples to a microarray that contains the tag compliments (Fig. 2). The development of this pooled fitness assay has allowed genome-wide screens to be performed in thousands of conditions, providing a wealth of functional information about the yeast genome (1–15). Pooled fitness assays have a variety of uses. Gene function can be studied by screening the collection in a condition whose impact on the cell is already known (Fig. 3a). For example, if a strain deleted for gene X is sensitive to a DNA damaging agent, it suggests that gene X is involved in DNA damage response (9,15, 18). This is a stronger test of gene function than examining changes in conditional expression because the phenotypic effect of each gene deletion is measured directly. This claim is supported by the observation that genes required for growth in a particular condition are often not overexpressed in that condition, and that a change in expression does not always indicate functional involvement (1, 4, 15, 16, 19). Pooled assays can also be used to study drug function. The function of a drug can often be determined by identifying the gene deletions that confer increased drug sensitivity or resistance

Fig. 1. Deletion cassette used for constructing strains in the yeast deletion collection. Each cassette carries the G418resistance marker KanMX4, which is required for the selection of transformants. The marker is flanked by two unique “barcode” or “tag” sequences, which are called the “uptag” and the “downtag.” These barcodes were designed to be maximally distinct, have uniform hybridization properties, and to carry no homology to the yeast genome (see Note 12). The two barcode sequences are flanked by four universal primer sites (arrows) that are common to all strains, and allow the barcodes to be amplified from a pooled culture. Uptags and downtags are amplified separately to avoid primer crosstalk. Integration into the genome is directed by two 45-bp homology regions which mediate replacement of the targeted gene by mitotic recombination (shown as black cross marks). Modified from ref. 38.

Chemogenomic Approaches to Elucidation of Gene Function

117

Fig. 2. Overview of the pooled fitness assay. Fitness profiling of pooled deletion strains involves five main steps: (1) Strains are first pooled at approximately equal abundance. (2) The pool is grown competitively in the condition of choice and a control condition. If a gene is sensitive to the treatment condition, the strain carrying this deletion will grow more slowly and become under-represented relative to the control culture (strain 3). Resistant strains will grow faster and become over-represented (not shown). (3) Genomic DNA is isolated from cells harvested at the end of pooled growth, and barcodes are amplified from the genomic DNA with universal primers. (4) PCR products are then hybridized to an array that detects the tag sequences, giving tag intensities for the two samples. (5) The treatment and control sample are then compared to determine the relative fitness of each strain. Note that only strain 3 is called as sensitive to the condition. While strain 2 grows more slowly than strain 1 in the treatment, this growth difference in not of interest because it matches that seen in the control. Modified from ref. 38.

because many genes in the yeast genome have annotated functions, making it likely that any given drug will perturb deletion strains with existing functional information (Fig. 3b). For example, if drug X perturbs deletion strains known to be sensitive to DNA damage, it suggests that drug X is a DNA damaging agent (13). Sensitivity data can also be used to identify subtle differences between compounds with similar modes of action, as well as secondary effects that may remain unknown even if the drug’s primary function is well characterized (9, 13, 16).

118

Pierce et al.

Fig. 3. Applications of homozygous deletion screening. Homozygous deletion screens can be used for studying gene function, drug function, and gene relationships. (a) Screening the homozygous deletion collection in a condition with a known impact on the cell provides functional information about the sensitive and resistant genes identified. (b) Screening a drug of unknown function will typically perturb some genes with existing functional annotations. These identified genes give information about the drug’s mode of action. (c) Genes that are functionally related will typically be sensitive in similar conditions. The combined results of many homozygous screens can therefore be used to give information about the functional relatedness of gene pairs. Functional relatedness is estimated by measuring the similarity of a gene-deletion’s response across many conditions.

In addition to generating information about single genes and single drugs, large sets of experiments can be used to study the relationships between genes and the relationships between drugs (Fig. 3c). For a given gene pair, the correlation of fitness values across conditions can be used to estimate the functional relatedness of the two genes, enabling large functional datasets to identify functional relationships between genes. Similarly, for a given drug pair the correlation of fitness values across genes is a measure of the similarity of their mechanism of action (15). Pooled analysis of the heterozygous deletion strains can be used to identify novel drug targets (Fig. 4) (6, 7, 10, 17). The reasoning behind this technique is as follows: if deletion of the drug target results in reduced growth, then heterozygous deletion of the gene encoding this target will often confer increased sensitivity to the drug. This is because the gene copy number can affect protein abundance (20), and if the heterozygote has a reduced level of the target protein, a lower dose of drug will be required to deplete enough of the protein’s activity to mimic the phenotype of the gene’s homozygous deletion, which is reduced growth. The heterozygote will therefore require a lower dose of drug than the wild-type strain to show the slow-growth phenotype that result from full inactivation of the target protein. This effect enables all genes whose full deletion confers growth sensitivity to be screened for possible drug targets in a single pooled experiment.

Chemogenomic Approaches to Elucidation of Gene Function

119

Fig. 4. Heterozygous profiling for drug target identification. Screening the heterozygous deletion collection can identify drug targets for genes whose homozygous deletion causes a growth defect. This method exploits the fact that a gene’s copy number is related to its expression level. Here a drug with a specific protein target is shown in black, and the corresponding protein target is shown as a circle. At an intermediate concentration, the drug will inactivate most target function in the heterozygote, but not in strains with two functional gene copies. This will result in a growth defect for the strain that carries a heterozygous deletion of the gene encoding the drug’s target. This phenotype mimics the growth defect caused by the corresponding homozygous deletion. Modified from ref. 38.

In addition to the applications mentioned above, these protocols can also be used to adapt the DNA barcode approach to other types of samples. Barcoded sample-tracking is a robust method that can be applied to a variety sample types. For example, DNA barcodes have already been used to tag mutant collections in other organisms (21, 22), collections of siRNA vectors (23–30), and DNA probes such as molecular inversion probes (MIPs) (31–34). The pooled fitness assay involves five main steps (Fig. 2, for an experimental timeline see Fig. 5). First, the deletion collection is grown on solid media in an arrayed format and the resulting cells are pooled and frozen to make cell aliquots that will be used for starting growth experiments. Cells are then grown in the desired conditions, typically for 5–20 generations. The barcodes from the resulting cell samples are prepared for hybridization by isolation of the genomic DNA and PCR amplification with the common barcode primers. Amplification is performed in two separate reactions to prevent crosstalk between the uptag and downtag primers. The resulting PCR products are then hybridized to tag microarrays. The barcode intensity data is then analyzed to determine differences in strain representation between pairs of samples.

120

Pierce et al.

Fig. 5. Timeline for a pooled growth experiment. (1) Pooled cells are generated by transferring a copy of the deletion collection from solid media to liquid and then freezing small aliquots. This step is time consuming, but is performed infrequently because many aliquots can be generated at once. (2) Cultures are inoculated using a frozen cell aliquot, and then grown for the desired number of generations. The specific amount of time needed for growth will vary depending on the number of generations, and on the growth rate of the culture in the treatment condition. The example shown is for 20 generations of growth either as (a) 700-ml cultures in a 48-well plate, shaken and grown in a spectrophotometer, or (b) 50-ml cultures grown in flasks. (3) Genomic DNA is purified from the harvested cells using a standard column-based purification kit, and tags are then PCR amplified from the purified genomic DNA. The uptags and downtags are amplified separately to avoid cross-reactions between the uptag and downtag primer pairs. (4) PCR products from both reactions are hybridized to a single array. After hybridization, the array is washed and scanned. The time needed for this step will vary depending on the hybridization time chosen. Modified from ref. 38.

2. Materials 2.1. Preparing Pooled Cell Aliquots

1. Frozen glycerol stocks of the yeast deletion collection in 96-well microtiter plates (OpenBiosystems, Part Nos. YSC1056 and YSC1055).

Chemogenomic Approaches to Elucidation of Gene Function

121

2. Nunc Omni Trays (VWR, Catalog No. 62409-600). 3. 96-Well pin tool (V&P Scientific, Catalog No. VP407A). 4. EtOH for flame-sterilizing pin tool (Gold Shield, Catalog No. DSP-CA-151). 5. Incubator set to 30°C for growing yeast plates. 6. Spectrophotometer. 7. 1,000× G418 Stock (200 mg/ml): Dissolve 5 g G418 (AgriBio, Catalog No. 3000) in 25 ml dH2O. Filter sterilize using a 0.2-mm filter and a syringe. Shield from light by wrapping bottle in foil. Store at 4°C. 8. YPD + 200 mg/ml G418 rectangular plates: Mix 10 g yeast extract (Becton Dickinson and Co., Catalog No. 212750), 20 g peptone (Becton Dickinson and Co., Catalog No. 211677), 20 g dextrose Becton Dickinson and Co., Catalog No. 215530), 20 g agar (Becton Dickinson and Co., Catalog No. 214010), and 1 l dH2O to a 2-l flask with a stir bar. Autoclave. Allow media to cool to approximately 50°C with gentle stirring. Add 1 ml 1,000× G418 Stock. Stir gently for an additional 1 min to ensure drug is evenly mixed. Pour into Nunc Omni trays, 50 ml per tray. Sufficient for approximately 20 plates. Store at 4°C. 9. YPD liquid + 200 mg/ml G418: Mix 10 g yeast extract (Becton Dickinson and Co., Catalog No. 212750), 20 g peptone (Becton Dickinson and Co., Catalog No. 211677), 20 g dextrose Becton Dickinson and Co., Catalog No. 215530), and 1 l dH2O to a 1-l bottle. Autoclave. Allow media to cool to approximately 50°C. Add 1 ml 1,000× G418 Stock. Store at 4°C. 2.2. Growing Pooled Cultures

1. Temperature controlled shaker for 250-ml flasks or spectrophotometer that allows incubation and shaking of plates, such as the TECAN GENIOS SpectraFLUOR Plus (Tecan, Part No. F129005, Note: many spectrophotometers do not shake plates hard enough for growing yeast). 2. 250-ml culture flasks (if growing cultures in flasks). 3. 48-Well plates (Greiner, Part No. 677102; if growing cultures in plates). 4. Plate roller (Sigma, Catalog No. R1275) for sealing 48-well culture plates (if growing cultures in 48-well plates). 5. Adhesive plate seals (ABgene, Catalog No. AB-0580) for sealing 48-well culture plates (if growing cultures in 48-well plates). 6. Desired media and compounds for growing cultures.

122

Pierce et al.

2.3. Purification and Amplification of Barcodes from Cell Samples

1. YeaStar Genomic DNA Kit (Zymo Research, Catalog No. D2002). 2. Taq DNA polymerase (Invitrogen, Catalog No. 10342). 3. dNTPs (Invitrogen, Catalog No. 10297). 4. 10× PCR reaction buffer (MgCl2) (Invitrogen, Part No. 52724). 5. 50 mM MgCl2 (Invitrogen, Part No. 52723). 6. Up primer mix: Dissolve Uptag (5¢ – GAT GTC CAC GAG GTC TCT – 3¢) and Buptagkanmx4 (5¢ biotin – GTC GAC CTG CAG CGT ACG – 3¢) each in dH2O at 100 pmol/ml, then mix in a 1:1 ratio for a final concentration of 50 pmol/ml each. Store at −20°C. 7. Down primer mix: Dissolve Dntag (5¢ – CGG TGT CGG TCT CGT AG – 3¢) and Bdntagkanmx4 (5¢ biotin – GAA AAC GAG CTC GAA TTC ATC G – 3¢) each in dH2O at 100 pmol/ml, then mix in a 1:1 ratio for a final concentration of 50 pmol/ml each. Store at −20°C. 8. Thermocycler with heated lid.

2.4. Array Hybridization and Scanning

1. Genflex Tag 16K Array v2 (Affymetrix, Part No. 511331). 2. Hybridization Oven 640 (Affymetrix, Part No. 800138). 3. GeneChip Fluidic Station 450 (Affymetrix, Part No. 00-0079). 4. GeneArray Scanner 3000 (Affymetrix, Part No. 00-0212). 5. 0.5-ml microfuge tubes suitable for boiling (Eppendorf 0.5ml Safe-Lock microcentrifuge tubes, Sigma, Catalog No. T8911). 6. Microcentrifuge Z708372).

tube

holders

(Sigma,

Catalog

No.

7. Ice bucket. 8. Boiling water bath with floating rack for 0.5-ml tubes. 9. Teeny Tough-Spots (Diversified Biotech, Catalog No. TSTNY; or any other small labels suitable for preventing evaporation from Affymetrix array gaskets). 10. Denhardt’s Solution, 50× Concentrate (e.g., Sigma, Catalog No. D-2532). 11. Streptavidin, R-phycoerythrin conjugate (SAPE) – 1 mg/ ml (Invitrogen, Catalog No. S-866). Store at 4°C. Do not freeze. 12. B213 oligonucleotide: Dissolve B213 control (5¢ biotin – CTGAACGGTAGCATCTTGAC – 3¢) in dH2O at 100 pm/ml to make a concentrated stock, and then use the stock to make 1 fmol/ml working solution. Store at −20°C.

Chemogenomic Approaches to Elucidation of Gene Function

123

13. Mixed oligonucleotides: Dissolve each of the following eight oligos in dH2O at 100 pmol/ml: Uptag (5¢ – GAT GTC CAC GAG GTC TCT – 3¢), Dntag (5¢ – CGG TGT CGG TCT CGT AG – 3¢), Uptagkanmx (5¢ – GTC GAC CTG CAG CGT ACG – 3¢), Dntagkanmx (5¢ – GAA AAC GAG CTC GAA TTC ATC G – 3’), Uptagcomp (5¢ – CTA CAG GTG CTC CAG AGA – 3¢), Dntagcomp (5¢ – GCC ACA GCC AGA GCA TC – 3¢), Upkancomp (5¢ – CAG CTG GAC GTC GCA TGC – 3¢), Dnkancomp (5¢ – CTT TTG CTC GAG CTT AAG TAG C – 3¢). Mix an equal volume of each of the eight oligos for a final concentration of 12.5 pmol/ml each. Note that the Uptag oligo is also used in Up primer mix and the Dntag oligo is also used in Down primer mix, so take care to leave enough for use in both mixes. Store at −20°C. 5 M NaCl Dissolve 58.4 g NaCl in 200 ml dH2O. Store at room temperature (18–25°C). 14. 0.5 M EDTA: Dissolve 7.3 g EDTA (BioRad, Catalog No. 161-0729) in 50 ml dH2O. Store at room temperature. 15. 10% Tween: Mix 2 ml Tween 20 (Sigma, Catalog No. T2700) with 18 ml dH2O. Store at 4°C. 16. 1% Tween: Mix 10 ml Tween 20 (Sigma, Catalog No. T2700) with 990 ml dH2O. Store at 4°C. 17. 12× MES stock: Mix 0.70 g MES free acid monohydrate (Sigma, Catalog No. M5287), 1.9 g MES Sodium Salt (Sigma, Catalog No. M5057), 8 ml of H2O and adjust volume to 10 ml. Adjust the pH to be between 6.5 and 6.7 using a pH meter or pH paper. Shield from light by wrapping bottle in foil. Store at 4°C. Replace if solution becomes visibly yellow or after 1 month. 18. 2× Hybridization buffer: Mix 8.3 ml of 12× MES Stock,17.7 ml of 5 M NaCl (J.T. Baker, Catalog No. 362401), 4.0 ml of 0.5 M EDTA, 0.1 ml of 10% Tween 20 (vol/ vol), 19.9 ml filtered dH2O. Filter to remove dust particles that would interfere with scanning. Shield from light by wrapping bottle in foil. Store at 4°C. Replace as for MES stock. 19. Wash A: Mix 300 ml 20× SSPE (Sigma, Catalog No. S2015), 1 ml 10% Tween (vol/vol), 699 ml filtered dH2O. Filter to remove dust particles that would interfere with scanning. Store at room temperature. 20. Wash B: Mix 150 ml 20× SSPE (Sigma, Catalog No. S2015), 1 ml 10% Tween (vol/vol), 849 ml dH2O. Filter to remove dust particles that would interfere with scanning. Store at room temperature.

124

Pierce et al.

3. Methods 3.1. Yeast Deletion Strain Pool Construction

1. Allow at least 1 week to generate pooled aliquots of cells before beginning pooled growth experiments. Pooling is performed infrequently because pooled cells can be stored indefinitely at −80°C. See Note 1 for information about strain background options. 2. Allow the frozen glycerol stocks for the strains of interest to thaw completely because cells may have settled prior to being frozen. Remove plates from the freezer in manageable numbers to avoid leaving all the plates at room temperature for an extended period. 3. To sterilize a 96-well pin tool, dip pin tool in water to rinse away any remaining cells, followed by two dips in 70% ethanol baths, then carefully flame the pin tool. Allow pin tool to cool for 1 min. Make certain that the level of the ethanol bath exceeds the level in the water bath to ensure all carry-over cells are flamed and removed. Change water frequently. 4. Insert the sterile 96-well pin tool into a thawed 96-well plate, swirl gently to capture settled cells and then transfer to a Nunc Omni Tray containing 50 ml of YPD-agar including 200 mg/ ml G418. Allow pin to dwell on agar for 5–10 s. Repeat for the remaining plates, sterilizing the pin tool between transfers. 5. Grow colonies until they reach maximal size at 30°C (2–3 days). 6. After colonies have reached full size, make note of any strains that are missing or appear as slow-growing colonies. Streak these strains out individually using standard yeast procedures to generate more cells. 7. Scrape the entire contents of all plates (in a laminar flow hood to avoid contamination) into a 50-ml conical centrifuge tube containing YPD liquid media + 200 mg/ml G418. Alternately, cells can be soaked off the plates by adding 12 ml YPD liquid media + 200 mg/ml G418 to each plate, soaking for 5 min, and gently agitating the plate to resuspend the cells. After cells are resuspended, pool liquid in a flask by transferring a fixed volume from each plate with a pipette. 8. For the slow-growing strains observed in step 5, add approximately three colony equivalents of cells using a sterile flat toothpick. This will help to ensure that these slow growing strains remain detectable by the end of a typical growth experiment. 9. Add glycerol to 15% or DMSO to 7% (vol/vol).

Chemogenomic Approaches to Elucidation of Gene Function

125

10. Measure the OD600 of the pool and adjust to a final concentration of 50 OD600/ml (OD600 1.0–2.2 × 107 cells/ml for diploid strains) with media containing 15% glycerol or 7% DMSO. If lower than 50 OD600/ml, centrifuge the pool at 500 × g for 2 min. to pellet the cells, decant the supernatant, and resuspend to 50 OD600/ml. 11. Aliquot 10–25 ml of pool into individually capped PCR tubes, and store at −80°C. Pooled aliquots can be stored indefinitely. 12. If desired, perform a hybridization test and create a pool supplement to boost the signal for strains that do not give good signal when hybridized (see Note 2). 3.2. Pooled Growth

1. Before beginning, see Subheading 4 for information about growth equipment options (see Note 3), using a pre-screen growth step to allow the strains to recover from freezing (see Note 4), choosing the number of generations of growth (see Note 5), selecting the correct dose for drug conditions (see Note 6), choosing a sufficiently large culture volume and starting OD (see Note 7), selecting the correct control condition (see Note 8), and replicating samples (see Note 9). 2. Thaw a frozen aliquot of pooled cells on ice. 3. Immediately dilute the pool into media containing conditions of interest, including a control condition. For drug experiments, if the drug added is dissolved in a solvent, an equal amount of solvent should be added to the control culture. Choose culture volumes, starting OD, and final OD using the guidelines in Note 7, or use one of the two options given below. (a) Option A: Use a starting OD600 of 0.0625 in two 700ml wells of a 48-well plate (see Note 10). Seal the plate with a plastic plate seal, rolling with a plate roller. If the condition requires aerobic growth (for example, nonfermentable carbon sources) poke a small hole in the membrane seal towards the side of each well to allow better aeration of the culture. Grow in a spectrophotometer at 30°C with shaking set to the highest setting. Cells can be grown in a single culture for five generations. (b) Option B: Use a starting OD600 of 0.0020 in a 50-ml culture in a 250-ml culture flask. Grow in a shaking incubator at 30°C and 250 rpm. Cells can be grown in a single culture for ten generations. 4. Grow cells for the desired number of generations, or until the OD600 = 2. If cells have reached the desired number of generations, proceed to the next step. Otherwise transfer the cells to a fresh culture as in the previous step.

126

Pierce et al.

5. Collect 1–2 OD600 of cells in a 1.5-ml tube. Spin the tubes at 500 g for 2 min in a microcentrifuge to pellet the cells, and remove the media. 6. If a starting cell sample (i.e., a “T0 time point”) is needed, add 1–2 OD600 of pool from the freezer to a 1.5-ml tube. Spin the tube at max for 2 min in a microcentrifuge and remove the media. 7. If not proceeding to the next step immediately, store cell pellets at −20°C in a non-frost free freezer. 3.3. Purification and Amplification of Barcodes from Cell Samples

1. Purify genomic DNA from the cell pellets with the Zymo Research YeaStar kit using Protocol I (included with the kit), or another suitable method for purifying yeast genomic DNA. Elute the DNA with 300 ml 0.1× TE. Genomic DNA can be stored indefinitely at −20°C. 2. If desired, quantitate genomic DNA using a gel or a UV spectrophotometer. In practice, using a consistent volume of genomic DNA (we use 15 ml which contains ~0.1 mg) from the previous step without adjusting for variations in DNA concentration gives good results. 3. Set up two 60 ml PCR reactions for each sample, one for the uptags and one for the downtags (33 ml dH2O, 6 ml 10× PCR buffer without MgCl2, 3 ml 50 mM MgCl2, 1.2 ml 10 mM dNTPs, 1.2 ml 50 mM Up or Down primer mix, 0.6 ml 5 U/ml Taq polymerase, ~0.1 mg genomic DNA in 15 ml). 4. Cycle as follows in a thermocycler with a heated lid: 94°C 3 min; repeat 30×: 94°C 30 s, 55°C 30 s, 72°C 30 s; 72°C 3 min, hold at 4°C. PCR products can be stored at −20°C. 5. If desired, check the resulting PCR products on a gel (see Note 11). 6. If not proceeding to the next step immediately, store PCR products at −20°C in a non-frost free freezer.

3.4. Array Hybridization and Scanning

1. These protocols use Affymetrix TAG4 arrays. See Note 13 for more array platform options. 2. Set up a boiling water bath with a floating rack for 0.5-ml tubes and a slushy ice bucket. Set hybridization oven temperature to 42°C. 3. Fill the arrays with 90 ml 1× hybridization buffer. 4. Incubate at 42°C and 20 rpm for at least 10 min in the hybridization oven to pre-wet the array. 5. Prepare 60 ml hybridization mix per sample (45 ml 2× hybridization buffer, 0.3 ml B213 control oligonucleotide (1 fm/ml) (see Note 14), 7.2 ml of mixed oligonucleotides (12.5 pm/ml) (see Note 15), 1.8 ml 50× Denhardt’s Solution) immediately before using and aliquot into 0.5-ml tubes suitable for boiling.

Chemogenomic Approaches to Elucidation of Gene Function

127

6. While arrays are equilibrating, add 20 ml of uptag PCR and 20 ml of downtag PCR to 60 ml of hybridization mix for a total volume of 100 ml. 7. Clip a tube-holder onto the lid of each tube to prevent the tubes from opening during boiling. 8. Boil each tube for 2 min and then set on ice for 2 min. 9. Spin tubes briefly to bring the samples to the bottom of the tubes. 10. Remove hybridization buffer from the arrays. 11. Add the samples to the arrays (90 ml per array). 12. Place a Tough-Spot over each of the two gaskets to prevent evaporation. 13. Hybridize for 3–16 h at 42°C and 20 rpm. It is best to keep the hybridization time constant for samples that are part of the same dataset. 14. Prime the fluidics station: put the tubings in place, empty the waste bottle, fill wash A, wash B, and water bottles, and run fluidics station protocol PRIME_450. 15. Prepare 600 ml biotin labeling mix per sample (180 ml 20× SSPE, 12 ml 50× Denhardt’s Solution, 6 ml 1% Tween 20 (vol/vol), 1 ml 1 mg/ml streptavidin–phycoerythrin, 401 ml dH2O). Prepare this mix immediately before using. 16. Aliquot 600 ml biotin labeling mix per chip into tubes. 17. Remove Tough-Spots from chips. 18. Remove hybridization mix and fill chips with 90 ml wash A. 19. Wrap chips that are waiting to be washed in aluminum foil. 20. Under Experiments: for each chip type in the sample name and the experiment name, enter the barcode. 21. Under Fluidics: enter the station, and which chip you will wash and label in which module using the “experiment” pull-down menu. Use the protocol “Genflex_TAG4_wash_ protocol.” 22. When the chips are ready for hybridization, check for air bubbles and wash again if necessary before you engage the wash block. 23. Clean the glass window on each array with isopropanol and a cotton swab or lint-free wipe. 24. Put Tough-Spots on the gaskets to prevent evaporation and put them in the scanner. 25. Scan at an emission wavelength of 560 nm. 26. When all fluidics operations are complete, put all cablings in Millipore water and run SHUTDOWN_450.

128

Pierce et al.

3.5. Analysis of Microarray Results 3.5.1. Outlier Masking

The Affymetrix TAG4 array contains at least five replicate features for each tag probe, dispersed across the array to make their performance as independent as possible (35). This allows outlier features to be identified and discarded before calculating an average intensity value for each tag (see Notes 12 and 16). 1. For each feature on the array, examine the features within the surrounding 5 feature × 5 feature region. If at least 13 of the 25 probes in this region differ from their trimmed replicate mean (the mean of the three middle replicates, excluding the highest and lowest replicates) by more than 10%, consider this probe part of an outlier-dense region that is not suitable for data analysis. 2. Once these outlier-dense regions have been identified, pad them by also including any probes that are within a 5-probe radius, as defined by ((x1 − x2)2 + (y1 − y2)2)1/2 < 6 where x1, x2, y1, and y2 are the x and y coordinates for the two features. 3. Also discard features for which (standard deviation of feature pixels/mean feature pixels) > 0.3. This unevenness is typically caused by bright pieces of debris on the chip that affect only a single probe, and are therefore not padded to include neighboring features. The standard deviation is included in the .cel file for Affymetrix arrays. 4. After identification and removal of outliers, intensity values are calculated for each tag by averaging all unmasked replicates.

3.5.2. Saturation Correction

Signal on the TAG4 array is not linearly related to tag concentration because of feature saturation (35). If left uncorrected, this saturation will cause the degree of sensitivity or resistance to be underestimated for strains with brighter tags. Saturation is corrected by comparing the uptag and downtag ratios (see Note 17). 1. Using a pair of arrays that are not sample replicates, calculate ln(ic − bg)/(it − bg) for each tag, where ic is the control intensity, it is the treatment intensity, and bg is the background as estimated by taking the mean intensity of the unassigned tag probes. 2. Mark the ratios for any tags with minimum values less than 3× background as unusable (this mark is for saturation correction only; a more robust threshold for unusable tags will be re-calculated later for use in the final results). 3. Pair uptag and downtag ratios by strain. Ignore ratios for any strains with less than two usable tags. 4. For each strain, calculate the difference in average intensity for the two tags: (itu + icu)/2 − (itd + icd)/2, where the subscript u indicates the uptag, and the subscript d indicates the downtag.

Chemogenomic Approaches to Elucidation of Gene Function

129

5. Sort ratios by the difference in average intensity as calculated above. Take a sliding window of 600 ratio pairs, sliding 100 pairs at a time. For each window fit a line to the uptag ratios (x-axis) versus the downtag ratios (y-axis) using least-squares and take the slope. Also calculate the mean of the differences in the average intensity for the window. 6. Fit a least-squares line to the intensity differences for all windows (x-axis), versus slopes for all windows (y-axis) and take the slope of this line. 7. Repeat steps 4–6 using the reverse intensity difference: (itd + icd)/2 − (itu + icu)/2, and taking the slope with the axes reversed: the downtag ratios on the x-axis, and the uptag ratios on the y-axis. 8. Average the slope calculated in step 6 with the slope calculated in the repeat of step 6. This is the saturation correction factor S. A typical range for S is 0.0001–0.0005. 9. Adjust the raw intensity data using the following transformation: f(i) = ie is. 10. To correct more than two arrays, calculate S for all possible pairs of arrays, then use the median of these values as the correction factor for all arrays in the set. Using a larger group of arrays will improve the accuracy of S. 3.5.3. Array Normalization

The uptags and downtags should be normalized separately, because they are amplified in separate PCR reactions, and the intensities of the individual PCR reactions will affect the intensities of their array. Normalize half of the unassigned probes with the uptags, and half with the downtags. These probes are normalized and saturation corrected with the other probes to keep their values comparable. Uptag background and downtag background are calculated separately, as normalization may have a different impact on the background level for each tag set. Normalize array data using either quantile normalization (option A) (36) or mean normalization (option B) (see Notes 18 and 19). 1. Option A: To quantile normalize a set of arrays: 2. For each set of tags (up and down), rank values obtained from each array in the order of increasing intensity. 3. For each rank, assign the tag at that rank for each array to the median of all values at that rank. 4. Option B: To mean normalize a set of arrays: 5. For each set of tags (up and down), divide by the mean. 6. Multiply each tag set by the mean across all arrays (this is for convenience only; it returns the tag intensities to approximately their original range).

130

Pierce et al.

3.5.4. Removing Unusable Tags

Tags with low intensity values in the control samples will give poor-quality results. An intensity value threshold for excluding these tags should be chosen by comparing the correlation of uptag and downtag ratios as a function of tag intensity (see Note 20). 1. Using any treatment-control pair, calculate log 2(ic − bg)/ (it − bg) for each tag, where ic is the control intensity, it is the treatment intensity, and bg is the mean intensity of the unassigned tag probes for the appropriate tag set (uptag or downtag). 2. Pair uptag and downtag ratios by strain. Ignore ratios for any strains with only one tag (~200 strains only have an uptag). 3. For each pair, take the minimum intensity for the two tags in the two samples. Sort the ratio pairs by this minimum intensity. 4. Use a sliding window of size 50 on the ranked ratio pairs, starting with the lowest intensity pairs. Calculate the correlation of uptag and downtag ratios for pairs within the window. Also calculate the average of the minimum intensities calculated in the previous step. 5. Slide the window by 25 pairs, and repeat the previous step until all pairs have been traversed. 6. Plot the average minimum intensity versus the uptag–downtag correlation for all windows. 7. Chose an intensity threshold from this plot by eye, or by using a set formula. For example, the intensity value where the correlation first reaches 80% of its maximum level is a good cutoff. 8. Mark any tags that are below this cutoff in either of the samples as unusable for calculating log ratios. Mark any tags below this cutoff in any of the control samples as unusable for calculating p-values. The usability criteria are less strict for p-value calculation than for log 2 ratio calculation because log 2 ratios are a measure of how sensitive each strain is, whereas p-values are only a measure whether the strain is sensitive or not. It is possible for a strain that is above background in the control samples but below background in the treatment sample to be significantly sensitive, but without a usable treatment sample the degree of sensitivity cannot be measured.

3.5.5. Calculating p-Values and False Discovery Rates

This method works best for cases where a large set of common control arrays (more than ten) are used with one or two replicate arrays per experiment (see Note 9). A validation of this method based on a comparison to the number of false sensitivity calls in matched control–control comparisons has been described in detail (15). If multiple replicates are available for the control and the treatment samples and a large control set is not available, other methods for calculating significance such as those used in the Significance Analysis of Microarrays (SAM) package (37) are preferable (see Notes 21–27 for more information about interpreting results).

Chemogenomic Approaches to Elucidation of Gene Function

131

1. For each usable tag, calculate the mean of the controls (mc) and the standard deviation of the controls (sc). 2. For each strain, average all usable tags. If a strain has no useable tags, this strain should be excluded from the analysis. 3. For each strain, calculate a z-score: (mc − t)/sc where t is the treatment intensity for that strain. Strains that are sensitive will have positive scores, while strains that are resistant will have negative scores. 4. Calculate p-values from the z-scores by fitting a t-distribution with nc − 1 degrees of freedom to all scores for the experiment, where nc is the number of control arrays. 5. Calculate a corresponding false discovery rate (FDR) for each of the p-values as follows. Rank the p-values in ascending order. For the p-value with rank r, the corresponding FDR is (pns)/r where ns is the total number of strains in the analysis. 3.5.6. Calculating log 2 Ratios

Whereas p-values describe the level of confidence for calling strain sensitivity, the log 2 ratios give the best estimate of the actual level of sensitivity for each strain. For example, deletions in the same complex or linear pathway will often cause similar growth defects, and as a result these strains will often have similar log 2 ratios. However, sensitivity for these strains may be measured with different degrees of accuracy, and they will therefore not necessarily have similar p-values (see Notes 21–27 for more information about interpreting results, and Note 28 for sample ratio results). 1. For each tag, calculate log 2(mc − bg)/(mt − bg) where mc is the mean intensity for the control samples, mt is the mean intensity for the treatment samples, and bg is the mean intensity of the unassigned probes for the appropriate tag set. 2. For each strain, average the log 2 ratios for all usable tags to obtain a final sensitivity score. 3. Strains that are sensitive will have positive scores, while strains that are resistant will have negative scores. This score is proportional to the log 2 ratio of cells present in the control sample versus the treatment sample.

4. Notes 1. Strain background options: The yeast deletion collection contains four different sets of strains: homozygous diploids, heterozygous diploids, MATa haploids, and MATa haploids. In general it is preferable to work with the heterozygous and the homozygous collections because the diploid genome is

132

Pierce et al.

more robust against phenotypic effects caused by secondary site mutations. These mutations are typically heterozygous in the diploid collections, whereas for the haploid collections a secondary site mutation will affect the only available gene copy. 2. Adjusting strains with low tag intensities: To identify strains that are underrepresented in the pool, perform a test hybridization. It is best to grow the pool in the standard control condition before this test hybridization, because slow growing strains often drop out of the pool during this growth step (15% of the homozygous deletion strains and 3% of the heterozygous strains are slow-growing when compared to wild type (1, 5)). Once identified, the signal for these strains can be boosted by adding extra cells. Strains that are not detectable can be added in larger quantities to a supplementary pool to boost their signal. It is normal that 200 strains or more remain undetectable even after interventions of this type. 3. Growth equipment options: There are many possible ways to grow pooled collections of deletion mutants. We grow our pooled cultures in shaking spectrophotometers that are connected to pipetting robots, allowing the cells to be inoculated into new cultures or harvested robotically when they reach the desired optical density (OD). This greatly facilitates the process and increases reproducibility by making the growth step as consistent as possible, however it is by no means necessary for obtaining good results. Cells can be grown in a shaking spectrophotometer to facilitate OD tracking and then transferred and harvested manually, or cells can be grown in flasks with both OD readings and transfers performed manually. Cells could also be grown as a lawn on solid media, or in alternate growth equipment such as a chemostat. 4. Using a pre-screen growth step to allow the strains to recover from freezing: Experimental cultures can either be inoculated directly from frozen aliquots or from a starting culture that has been grown for a set number of generations before starting the experiment. Including a recovery step allows the cells to recover from freezing, and may improve the accuracy of the sensitivity results. The main disadvantage to including this step is that increasing the total growth time causes more slowgrowing strains to drop out of the pool by the final time-point, resulting in a greater number of strains with unusable data. 5. Choosing the number of generations of growth: In general, a longer growth period can more sensitively detect subtle phenotypes, but results in the depletion of slow-growing strains from the population (in both the experimental and control condition), thus precluding the analysis of these strains. Therefore, the ideal growth period for a pooled culture will

Chemogenomic Approaches to Elucidation of Gene Function

133

vary. For example, it is generally preferable to grow heterozygous strains longer than homozygous strains because heterozygous growth phenotypes are generally more subtle. For resolution across a wide array of sensitivities, cells can also be collected at multiple time-points. 6. Selecting the correct dose for drug conditions: In general, a stronger treatment dose will reveal more sensitive strains by making the inhibition of slightly sensitive strains easier to detect. However, a higher dose will also increase the number of strains that do not grow at all, precluding the ranking of these strains’ sensitivities. In general, we aim for treatment doses that cause a 10–20% decrease in growth rate when tested on a wild type. Standardizing the level of inhibition allows most experiments to be sampled after the same number of generations of growth. At 10–20% inhibition, optimal results for the heterozygous collection are usually obtained at the 20 generation time point, and optimal results for the homozygous collection are usually obtained at the five generation time point. In general, a more benign treatment will cause smaller changes in growth rate, and will therefore require a longer growth period to resolve differences between strains. 7. Choosing a sufficiently large culture volume and starting OD: Choosing an appropriate culture volume is critical for obtaining good results. At each inoculation step, the cells are diluted to their lowest numbers. Decreasing the average number of cells per strain at inoculation increases the effect of variation on the actual number of cells sampled for each strain. This variation is typically the most significant source of noise in the assay (38). Using at least 300 cells per strain is recommended. The number of cells per strain at inoculation can be calculated as follows: Cells per strain = (Inoculation OD) × (culture volume (ml)) × (cells/ml at OD = 1)/(number of strains). A culture of diploid cells at OD600 1.0 has approximately 2.2 × 107 cells/ml. If the culture vessel will not accommodate the volume needed to reach the desired cell numbers, multiple cultures can be grown in parallel and pooled at the end of the experiment. Increasing the number of inoculations will also increases noise, and should be minimized. Also keep in mind the specific needs of the growth setup when choosing a culture volume (for example, manual OD readings will require extra culture volume). 8. Selecting the correct control condition: Raw values of array intensity for a single sample are very poorly correlated with the abundance in absolute strain. For this reason, each experiment requires at least two samples – a control sample and a treatment sample. The intensity ratio between two samples is a good measure of the abundance ratio of the corresponding strain (35,

134

Pierce et al.

38). The appropriate control condition will vary depending on the experiment. To measure relative growth rates, different time-points should be compared. To measure the relative sensitivity of each strain to a treatment condition, cells should be grown in a treatment condition and a control condition for the same number of generations. This type of comparison is useful because it takes each strain’s normal growth-rate into account, preventing strains that are slow-growing in the absence of treatment, from appearing falsely sensitive. 9. Sample replicates: Each experiment requires at least two arrays – one for a control sample, and one for a treatment sample. A single treatment-control pair gives good-quality data, especially for cases where the data set taken as a whole is more important than the accuracy of any single experiment. For example, in cases where the results of many experiments are used to build a gene network, the results will be relatively robust against false positives, and testing more conditions may improve the results more than increasing the number of experimental replicates. For cases where individual gene sensitivity predictions are important, replication may also be unnecessary if array experiments are confirmed with singlestrain follow up tests. For more accurate data, two or more replicates of both the control and treatment conditions are recommended to allow for the identification and removal of rare outlier points. Increasing the number of replicates will also allow for more subtle phenotypes to be detected with statistical significance. We often use a large, shared set of control arrays (more than ten) for analyzing many different experimental arrays, each with only one replicate. This control set can be used to calculate the statistical significance of the final results, while minimizing the total number of experimental arrays needed. If using a large, shared control set it is important that the control arrays represent as diverse a set of samples as the treatment arrays (cells grown on different days, tags amplified in different runs, etc.) so that the variation in the control set will accurately represent the nontreatment related variation present in the treatment arrays. When not using a shared set of control arrays, it is best for each control-treatment pair to be processed together (cells grown on the same day, tags amplified in the same run, etc.) to minimize the non-treatment related differences between the two. As discussed in Note 7, the growth step is the main source of noise in the assay. For this reason, all replicate samples should be full biological replicates (grown separately, as opposed to separate hybridizations of the same growth sample). It is important to note that the uptag and downtag are not independent measures of strain sensitivity. The two tags will share the noise generated during the growth step, where any sampling error affects both tags simultaneously.

Chemogenomic Approaches to Elucidation of Gene Function

135

The uptag and downtag therefore can not be treated as independent sample replicates when using statistical analysis methods. All useable tags for each strain should therefore be averaged before using these methods, with only true experimental replicates treated as replicate samples. 10. Comparing OD600values between plates and cuvettes: If growing cells in a shaking spectrophotometer, note that the OD600 measured for the plate will differ from the OD600 of the same culture measured in a cuvette due to differences in path length. Similarly, OD600 readings in a shaking spectrophotometer will also vary with differences in culture volume. All OD600 values given in this paper refer to those measured in a 1 ml, 1-cm path length cuvette. 11. Checking for successful tag amplification: To check for successful amplification, tagged PCR products can be separated on a gel. The desired product is ~60 bp. A second band is often seen because non-complementary tags can hybridize at their common primer regions, forming a partially single-stranded structure that migrates more slowly than the fully double-stranded tag products. Reactions that appear unusually dim (less than 5× the concentration of a typical reaction) will often give low-quality array data. These amplifications can be repeated using the leftover genomic DNA. It is common to have amplification in no-template control reactions because tag contamination is extremely difficult to prevent. This low-level of contamination will typically not impact results because the concentration of the experimental template will be much greater than the concentration of contamination in the experimental samples, preventing the contamination from having a serious impact on the results. 12. Barcode sequences: The deletion of a given gene will use the same two barcodes in all four strain backgrounds. Because of this barcode overlap, the homozygous deletion collection and the heterozygous deletion collection must be pooled and grown separately to ensure that each barcode corresponds to a unique strain within the pool. A full list of sequences is available (38). It is an Affymetrix convention to list probe sequences for their arrays in 3¢®5¢ orientation because this is the direction of probe synthesis used in their manufacturing process. For this reason, the tag sequences listed in Affymetrix array files are the reverse of those listed elsewhere. This is a common source of confusion, so sequence orientation should be carefully checked when designing new barcoded strains or other samples for use with Affymetrix tag arrays. 13. Array options: These protocols use the Affymetrix tag arrays; however they can easily be adapted to the array platform of your choice. For alternative array examples see refs. 4, 12, 13, 39, 40.

136

Pierce et al.

14. Determining the correct B213 concentration: It is common when diluting a new batch of B213 control oligo to have to test several concentrations to find a concentration that gives good results. The reason for this inconsistent performance from batch to batch is unknown. The B213 oligo binds to the border probes on the array, which are used by the Affymetrix software to align the grid that defines the probe borders. If the control spots are too dim, automated alignment will fail, and if the control probes are too bright, they can interfere with neighboring probes on the array. It is best to aim for an intensity that is bright enough for alignment to work consistently, but within the range of the tag probe intensities. If grid alignment fails because the border probes are too dim, it is possible to align the grid manually and recalculate the probe intensities using the Affymetrix software. 15. Mixed oligos: The mixed oligos are added to prevent re-annealing of the tagged product after melting. Non-complementary tags can stick together by their common primer sequences, forming a partially single-stranded product. Oligos complementary to the common primers are added in excess to prevent the formation of this product, making the tag sequences more accessible for hybridization. 16. Re-hybridizing failed samples: The PCR volume recommended here provides enough PCR for two hybridizations. If the first hybridization has problems such as an extremely uneven appearance (for example, due to large bubbles during washing or staining) or an unusually large amount of visible debris or damage, the remaining PCR can be hybridized to a new array. The replicate features can correct for minor damage, so in general, only damage that affects more than approximately one fifth of the array surface is a concern. 17. Saturation correction: The saturation correction function is derived from the data by exploiting the fact that the raw intensities for a strain’s uptag and downtag often differ, despite the fact that each uptag–downtag pair is measuring the same true strain ratio. The saturation pattern can therefore be derived by comparing the difference in raw intensity to the difference in ratio for all uptag–downtag pairs. The data is then corrected to make the ratios independent of raw intensity. This transformation increases the correlation of uptag and downtag ratios. The correction algorithm requires at least two arrays, and works best with a pair of arrays that are not sample replicates. This is because all log ratios for a pair of sample replicates will be approximately zero, and saturation affects ratios close to zero the least. The level of saturation can vary from day to day because of differences in hybridization time, washing, and staining time. For this

Chemogenomic Approaches to Elucidation of Gene Function

137

reason it is best to derive the correction function for all arrays hybridized on a given day as a group, and avoid using pairs of arrays processed on different days for deriving the correction function. 18. Normalizing array data: Two alternate normalization methods are discussed here – mean normalization, and quantile normalization (36). Quantile normalization is often preferred because it can normalize data effectively even when there is a non-linear relationship between the samples, but this method requires the assumption that the overall distribution of tag intensities is the same for all experiments being normalized. Mean normalization, though only capable of linear transformation of the data, does not require this assumption. In general, quantile normalization works best for experiments in which a reasonably small number of strains are affected by the condition tested. For experiments with a large number of sensitive strains, mean normalization often performs better. In particular, experiments with different growth periods (for example, the zero time-point and the 20 generation time-point) or from different starting pools (for example, homozygous pools constructed separately) should not be normalized together using quantile normalization because the tag intensities for these samples will have very different distributions (38). 19. Normalizing new arrays without renormalizing old data: A slight variation on the normalization procedures described can help make data analysis more convenient by allowing new data to be normalized without having to re-normalize all existing data. When quantile normalizing a group of arrays, a standard curve is usually derived from the all arrays in the group. If you wish to keep the standard curve constant over time, it can be calculated from a starting set of arrays and then held constant for normalizing future arrays. This works well as long as all arrays normalized are of the same experimental type (same pooled cells, grown for the same number of generations). Similarly, when mean normalizing a set of arrays, the mean intensity across all arrays in the set is used to return the mean-normalized values to their original range. This mean can also be calculated from an initial set of arrays and then held constant. 20. Removing unusable tags: When tags become too dim to be usable, their correlation with their tag replicates degrades to zero. By comparing the correlation of uptag and downtag ratios to tag intensity, an effective cutoff can be chosen to remove these tags. In practice, this intensity cutoff remains relatively constant across experiments as long as the experimental setup (in particular the hybridization time and the scanner used) does not change, so once a threshold is chosen, it is not required to be recalculated with every experiment.

138

Pierce et al.

21. Reasons expected hits may be absent: If a strain for which a sensitive phenotype was expected is not called as significantly sensitive, a good first thing to check is what the maximum achievable score is for that strain based on the control experiments. For example, if a strain is already fairly close to the background in the controls, or if the control data for that strain is noisy, it may not be possible for the experimental sample to be dim enough to give a significant score. To calculate the maximum achievable score for each strain, create a mock experimental chip in which all of the tag values are set to background, which is the minimum possible experimental value for each strain. Calculate scores using this mock-experimental chip and the control set, and apply the same FDR cutoff used for the true experimental results. This will give the maximum score achievable for each strain. If a strain is not called as sensitive in this mock experiment, then the data for that strain can be considered inconclusive. It is also possible that the pooled growth phenotype for the strain of interest differs from its phenotype in another environment. Growth in a shaking spectrophotometer, growth in a flask, and growth on a plate may be slightly different because of differences in these growth environments. For example, a culture in a shaking spectrophotometer typically has a lower oxygen level than a standard flask culture. Pooled growth may also create differences in phenotype. A deletion strain grown in a pooled culture is predominantly surrounded by strains that are wild type at its deleted locus. Under certain conditions these surrounding strains may be able to compensate for the deficiencies caused by the deletion. 22. Evaluating data quality and replicating sample agreement: The most effective way to measure the quality of replicate samples is to measure the correlation of the log-transformed, normalized, and saturation-corrected intensity values, termed “processed raw values.” This correlation should always be at least 0.98 for replicate samples. Processed raw values are a better indicator of sample quality than raw values because normalization and saturation correction remove artifacts that will artificially decrease replicate correlation, and log transformation adjusts the values so that a given deviation from the diagonal will have a consistent impact on the final log ratios, regardless of tag intensity. Lower correlation may indicate a sampling error problem, which can be corrected with increased culture volume as described in Note 7. Uptag and downtag ratios for a given sample will be positively correlated; however it is important to note that low correlation for uptag and downtag ratios does not necessarily indicate poor sample quality. In experimental samples where few strains are sensitive, most of the ratios are centered at zero, causing low correlation. In perfect data, the uptag–downtag log 2 ratio correlation for a control–control pair would be

Chemogenomic Approaches to Elucidation of Gene Function

139

zero, but sampling error causes slight strain representation differences between replicates, and the two tags correlate in measuring this noise. Replicate samples for which few strains are sensitive will also have low log 2 ratio correlation for the same reason. Because of this property, log 2 ratio correlation is a poor measure of replicate sample agreement. 23. Comparing results across different treatment conditions: The same score in two different experiments does not indicate the same level of inhibition relative to the control culture. To understand why, note that for any given condition, the phenotype for the majority of gene-deletion mutants is not significantly different from that of the wild-type control strain, and the median score for the experiment can therefore be assumed to be approximately that of the wild-type strain. This median score is also typically close to zero, because the final sample always contains the same number of cells no matter how slow growing a culture is. Therefore, for an experiment where the treatment culture grew 20% slower than the control culture, strains with a log 2 ratio of zero also grew approximately 20% slower than in the control culture. Similarly, in a case where the treatment culture grew 15% slower than the control culture, strains with a log 2 ratio of zero grew approximately 15% slower than in the control culture. Therefore, when comparing treatment inhibition for the same strain across multiple treatments, the growth rate of each pooled culture needs to be taken into account. Theoretically it should be possible to adjust log 2 ratios by multiplying the scores for each experiment by the average level of inhibition for each culture, however this idea has not been tested in practice. 24. Reduced data quality for slow growing strains: The data for strains with lower-than-average representation in the pool is noisier due to increased sampling error. Strains that are slowgrowing in the control condition are especially prone to this problem because they often have low representation in the pool (5). Data from these strains should be carefully confirmed. It is also common for slow-growing strains to appear slightly resistant to treatments that cause a large decrease in growth rate. While the reason for this effect is unknown, it is thought that slow growth may be less limiting in a condition where the pool as a whole is growing more slowly, causing these strains to be slowed proportionately less than others and thereby appear resistant. 25. Frequently sensitive strains: Some gene deletions are known to cause sensitivity in a large number treatment conditions (13, 15). While the sensitivity of these deletions is genuine, strains specifically sensitive to the condition of interest may

140

Pierce et al.

be better candidates for follow-up experiments than genes known to mediate sensitivity to a broad range of conditions. Common gene families include: ergosterol biosynthesis, aromatic amino-acid biosynthesis, genes related to transcription, and genes with a variety of membrane functions such as ER to Golgi transport, vesicle mediated transport, and vacuolar targeting (5, 15). The frequency of sensitivity for each strain in a large number of published experiments can be used as a reference when evaluating results (15). 26. Interference from neighboring gene deletions: Gene function can be disrupted by deletion of a neighboring gene, either by partial deletion of the ORF due to overlap, disruption of promoter regions, or by generation of secondary site mutations during homologous recombination. It is common to see pairs of adjacent genes that are both sensitive to the same conditions. The true gene responsible for the phenotype can be determined by adding each of the two genes back to each of the two deletion strains on a low copy-number plasmid. The gene that is responsible for the phenotype should complement the deletion in both of these strains. 27. Interpreting data for drug-target identification: In addition to the deletion strain corresponding to the drug target, there may be other heterozygous deletions that are sensitive. For example, a gene involved in coping with the stress caused by the drug may also be sensitive due to reduced protein level. Because of this, there is no set way for determining which, if any, of the sensitive strains corresponds to the drug target from the heterozygous profiling data alone. Other data types are often helpful for narrowing the candidates. Synthetic lethality data can aid in interpreting the results (12, 13, 39, 41, 42). Homozygous deletions corresponding to synthetic lethal partners of the drug target should be more sensitive to the drug because the drug mimics a deletion (or partial deletion) of the drug target. Synthetic lethality data is available for many complete deletions and also for essential genes in the form of promoter replacement alleles (43). and 3¢ untranslated region disruption alleles (44). both of which reduce gene dosage. Both of these data types are useful for clarifying heterozygous profiling results. Genetic interaction data can be found in the BioGRID database (45). Genome-wide overexpression data is also useful, as drug resistance often results from overexpression of the gene target (17). In addition, the corresponding homozygous deletion data (if a full deletion of the gene of interest is viable) should be examined. If both the heterozygote and the homozygote are sensitive the gene is unlikely to encode the drug target because a strain that lacks the drug target should not be affected by the drug.

Chemogenomic Approaches to Elucidation of Gene Function

141

28. Sample results: Fig. 6 shows sample data for an experiment in which the amino acid lysine was omitted. This data shows the typical level of agreement for raw data from multiple experiments, and the typical transformation of the raw data caused

Fig. 6. Sample data showing data analysis steps and expected results. Sample data is shown for a control–control pair, and for a lysine drop-out treatment sample compared to a control. (a) Raw tag intensities for the control–control pair (replicate array features have been averaged for each tag). (b) Data after saturation correction. Note that the range of intensities increases. (c) Data after quantile normalization. Note that the data is now centered along the diagonal. (d–f) The same steps are shown for the lysine drop-out treatment versus a control. (g) log 2 intensity ratios for the lysine drop out experiment. Positive values indicate sensitivity in the treatment condition, negative values indicate resistance. As expected, genes known to be involved in lysine biosynthesis are overrepresented among the sensitive strains (p = 1.13 × 10−14). Modified from ref. 38.

142

Pierce et al.

by saturation correction and quantile normalization. Data is shown for both a pair of controls and a control-treatment pair. As expected, the raw data for the pair of control samples is more similar than the raw data for the control-treatment pair. Final log 2 ratios are shown for the lysine drop-out condition versus the control condition. As expected, the majority of sensitive strains have known roles in lysine biosynthesis. Note that the log 2 ratios are similar for these strains, suggesting that they have a similar level of sensitivity.

Acknowledgments Work In the Giaever and Nislow labs is supported by the NHGRI, the Canadian Foundation for Innovation, and the CIHR (MOP81340 to GG) and (MOP-84305 to CN). References 1. Giaever, G. et al. Functional profiling of the Saccharomyces cerevisiae genome. Nature 418, 387–91 (2002). 2. Winzeler, E.A. et al. Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science 285, 901–6 (1999). 3. Shoemaker, D.D., Lashkari, D.A., Morris, D., Mittmann, M. & Davis, R.W. Quantitative phenotypic analysis of yeast deletion mutants using a highly parallel molecular bar-coding strategy. Nat Genet 14, 450–6 (1996). 4. Birrell, G.W. et al. Transcriptional response of Saccharomyces cerevisiae to DNA-damaging agents does not identify the genes that protect against these agents. Proc Natl Acad Sci USA 99, 8778–83 (2002). 5. Deutschbauer, A.M. et al. Mechanisms of haploinsufficiency revealed by genome-wide profiling in yeast. Genetics 169, 1915–25 (2005). 6. Giaever, G. et al. Chemogenomic profiling: identifying the functional interactions of small molecules in yeast. Proc Natl Acad Sci USA 101, 793–8 (2004). 7. Giaever, G. et al. Genomic profiling of drug sensitivities via induced haploinsufficiency. Nat Genet 21, 278–83 (1999). 8. Kastenmayer, J.P. et al. Functional genomics of genes with small open reading frames (sORFs) in S. cerevisiae. Genome Res 16, 365–73 (2006).

9. Lee, W. et al. Genome-wide requirements for resistance to functionally distinct DNA-damaging agents. PLoS Genet 1, e24 (2005). 10. Lum, P.Y. et al. Discovering modes of action for therapeutic compounds using a genomewide screen of yeast heterozygotes. Cell 116, 121–37 (2004). 11. Ooi, S.L., Shoemaker, D.D. & Boeke, J.D. A DNA microarray-based genetic screen for nonhomologous end-joining mutants in Saccharomyces cerevisiae. Science 294, 2552–6 (2001). 12. Parsons, A.B. et al. Integration of chemicalgenetic and genetic interaction data links bioactive compounds to cellular target pathways. Nat Biotechnol 22, 62–9 (2004). 13. Parsons, A.B. et al. Exploring the mode-of-action of bioactive compounds by chemical-genetic profiling in yeast. Cell 126, 611–25 (2006). 14. Steinmetz, L.M. et al. Systematic screen for human disease genes in yeast. Nat Genet 31, 400–4 (2002). 15. Hillenmeyer, M.E. et al. The chemical genomic portrait of yeast: uncovering a phenotype for all genes. Science 320, 362–5 (2008). 16. Hoon, S. et al. An integrated platform of genome-wide assays reveals small molecule bioactivities. Nat Chem Biol 4(8), 498–506 (2008). 17. Ericson, E. et al. Off-target effects of psychoactive drugs revealed by genome-wide assays in yeast. PLoS Genet 4(8), e1000151 (2008).

Chemogenomic Approaches to Elucidation of Gene Function 18. Workman, C.T. et al. A systems approach to mapping DNA damage response pathways. Science 312, 1054–9 (2006). 19. Jensen, L.J., Jensen, T.S., de Lichtenberg, U., Brunak, S. & Bork, P. Co-evolution of transcriptional and post-translational cell-cycle regulation. Nature 443, 594–7 (2006). 20. Pollack, J.R. et al. Microarray analysis reveals a major direct role of DNA copy number alteration in the transcriptional program of human breast tumors. Proc Natl Acad Sci USA 99, 12963–8 (2002). 21. Groh, J.L., Luo, Q., Ballard, J.D. & Krumholz, L.R. A method adapting microarray technology for signature-tagged mutagenesis of Desulfovibrio desulfuricans G20 and Shewanella oneidensis MR-1 in anaerobic sediment survival experiments. Appl Environ Microbiol 71, 7064–74 (2005). 22. Karlyshev, A.V. et al. Application of high-density array-based signature-tagged mutagenesis to discover novel Yersinia virulence-associated genes. Infect Immun 69, 7810–9 (2001). 23. Berns, K. et al. A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 428, 431–7 (2004). 24. Brummelkamp, T.R. et al. An shRNA barcode screen provides insight into cancer cell vulnerability to MDM2 inhibitors. Nat Chem Biol 2, 202–6 (2006). 25. Fischer, K.D. et al. Defective T-cell receptor signalling and positive selection of Vav-deficient CD4 + CD8 + thymocytes. Nature 374, 474–7 (1995). 26. Fraser, A. RNA interference: human genes hit the big screen. Nature 428, 375–8 (2004). 27. Kolfschoten, I.G. et al. A genetic screen identifies PITX1 as a suppressor of RAS activity and tumorigenicity. Cell 121, 849–58 (2005). 28. Ngo, V.N. et al. A loss-of-function RNA interference screen for molecular targets in cancer. Nature 441, 106–10 (2006). 29. Ngo, V.N. et al. A loss-of-function RNA interference screen for molecular targets in cancer. Nature 441, 106–10 (2006). 30. Westbrook, T.F. et al. A genetic screen for candidate tumor suppressors identifies REST. Cell 121, 837–48 (2005). 31. Akhras, M.S. et al. PathogenMip assay: a multiplex pathogen detection assay. PLoS ONE 2, e223 (2007).

143

32. Clayton, D.G. et al. Population structure, differential bias and genomic control in a largescale, case-control association study. Nat Genet 37, 1243–6 (2005). 33. Hardenbol, P. et al. Multiplexed genotyping with sequence-tagged molecular inversion probes. Nat Biotechnol 21, 673–8 (2003). 34. Hardenbol, P. et al. Highly multiplexed molecular inversion probe genotyping: over 10,000 targeted SNPs genotyped in a single tube assay. Genome Res 15 , 269–75 (2005). 35. Pierce, S.E. et al. A unique and universal molecular barcode array. Nat Methods 3, 601–3 (2006). 36. Bolstad, B.M., Irizarry, R.A., Astrand, M. & Speed, T.P. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19, 185–93 (2003). 37. Tusher, V.G., Tibshirani, R. & Chu, G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98, 5116–21 (2001). 38. Pierce, S.E., Davis, R.W., Nislow, C., & Giaever, G. Genome-wide analysis of barcoded S. cerevisiae gene-deletion mutants in pooled cultures. Nat Protoc 2, 2958–74 (2007). 39. Pan, X. et al. A robust toolkit for functional profiling of the yeast genome. Mol Cell 16, 487–96 (2004). 40. Yuan, D.S. et al. Improved microarray methods for profiling the Yeast Knockout strain collection. Nucleic Acids Res 33, e103 (2005). 41. Tong, A.H. et al. Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science 294, 2364–8 (2001). 42. Tong, A.H. et al. Global mapping of the yeast genetic interaction network. Science 303, 808–13 (2004). 43. Davierwala, A.P. et al. The synthetic genetic interaction spectrum of essential genes. Nat Genet 37, 1147–52 (2005). 44. Schuldiner, M. et al. Exploration of the function and organization of the yeast early secretory pathway through an epistatic miniarray profile. Cell 123, 507–19 (2005). 45. Stark, C. et al. BioGRID: a general repository for interaction datasets. Nucleic Acids Res 34, D535–9 (2006).

Chapter 8 Identification of Inhibitors of Chromatin Modifying Enzymes Using the Yeast Phenotypic Screens Benjamin Newcomb and Antonio Bedalov Summary A multitude of enzymes that modify histones and remodel nucleosomes are required for the formation, maintenance, and propagation of the transcriptionally repressed chromatin state in eukaryotes. Robust phenotypic screens in yeast S. cerevisiae have proved instrumental in identifying these activities and for providing mechanistic insights into epigenetic regulation. These phenotypic assays, amenable for high throughput small molecule screening, enable identification and characterization of inhibitors of chromatin modifying enzymes largely bypassing traditional biochemical approaches. Key words: Chromatin, Chemical genetics, Histone deacetylase, Chromatin modifying enzyme, Epigenetics

1. Introduction 1.1. Understanding Chromatin Biology Through Chemical Genetics 1.1.1. Chromatin Structure and Known Chemical Modifiers

Portions of the genome in eukaryotes is maintained in a transcriptionally inactive, or silenced state as a result of the local chromatin structure. The formation and the maintenance of the silenced state is an active process requiring a multitude of enzymes that act on DNA (methylation) and histones (acetylation, phosphorylation, and methylation) (1). Following cell division the silenced state of chromatin is passed onto the daughter cells thus forming a basis for the epigenetic propagation of cellular memory. Faithful transmission of the epigenetic state from mother to daughter plays a key role in many cellular processes in eukaryotes such as mating in yeast (2) or development in multicellular organisms (3). Epigenetic mechanisms also play an important role in the pathogenesis of many human neoplasms (4). The importance

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_8, © Humana Press, a part of Springer Science + Business Media, LLC 2009

145

146

Newcomb and Bedalov

of epigenetic regulation in cancer is underscored by the observations that tumor suppressor genes are often silenced rather than mutated and that many dominant oncogenes require epigenetic regulators for their activity. These epigenetic underpinnings of cancer can be exploited as a therapeutic strategy for two reasons. First, since silenced copies of tumor suppressor genes do not harbor genetic mutations, their reactivation in the context of malignant cells may suppress growth or induce death. Second, while transcription factors have traditionally been considered poor drug targets, the enzymatic activities required for their function (e.g., histone acetyl transferases HAT, histone deacetylases HDAC) can be inhibited pharmacologically. Together, these observations point to epigenetic regulation as a major new therapeutic area for cancer. At the present time our ability to pharmacologically influence epigenetics in cancer cells, and to use this as therapy, is limited by the scarcity of effective small molecule inhibitors of enzymes that control epigenetic states. Classic HDAC inhibitors (e.g., SAHA) and DNA demethylating agents (e.g., deoxy-5-azacytidine) are the only two classes of chromatin modifying drugs in clinical use. This highlights the need to develop new drugs that target other enzymes involved in the establishment and maintenance of epigenetic states. Traditional approaches to identify enzyme inhibitors rely on high throughput biochemical screens. However, the enzymatic activities and proteins required for epigenetic regulation are extremely well conserved among eukaryotes, which makes drug discovery possible in vivo using model organisms. Yeast is an attractive model system because of its rapid growth rate, ease of genetic manipulation, and because many yeast strains have already been developed to study epigenetics. Using a cell-based screen for compounds that can abrogate silencing at telomeres in yeast we have identified splitomicin, the first inhibitor of Sir2, a major nuclear NAD-dependent histone deacetylase and epigenetic regulator in yeast (5) and a founding member of a broadly conserved class of enzymes, sirtuins (6). Conditional inactivation of Sir2 with splitomicin and its analogues has proved valuable in dissecting chromatin biology in yeast (5, 7, 8) and mammalian cells (9), and in evaluating inhibition of sirtuins as a therapeutic strategy in cancer (10). Our success in identifying Sir2 inhibitors through phenotypic screens for epigenetic regulators in yeast, suggests that the same strategy can be used for the identification of inhibitors of other enzymes required for propagation of epigenetic memory. In the following sections we provide an overview of silencing in yeast, the enzymatic activities required for efficient silencing, and a description of the silencing assays available. Additionally, we provide a detailed high throughput screening protocol for identifying compounds that disrupt telomeric silencing, a description of the methods employed for

Identification of Inhibitors of Chromatin Modifying Enzymes

147

characterization of the hits, and an overview of the strategies for identifying the molecular targets of the compounds. 1.1.2. Yeast Silent Chromatin and Enzymatic Activities

Silent chromatin occurs at three distinct sites in the yeast genome: silent mating-type loci (HML and HMR), telomeres and at ribosomal RNA genes (rDNA) (11). The formation of silent chromatin, best understood at the silent mating type loci and telomeres, depends on DNA elements or silencers. These silencers are located in close proximity to the genes they regulate and contain binding sites for several DNA binding proteins including Rap1, Abf1, and the origin recognition complex. These DNA binding proteins, through protein–protein interactions, recruit the SIR (silent information regulator) complex (Sir2–4). Once nucleated at the silencers, the SIR complex is thought to spread along the chromatin through the binding of Sir3 and Sir4 to the hypoacetylated NH2-terminal tails of histones H3 and H4. The NH2-terminal tails of histones H3 and H4 are kept in the hypoacetylated state through the action of Sir2 and this activity is critical for the formation of silent chromatin. Sir2 deacetylase activity is also required for silent chromatin formation at rDNA. However, at rDNA Sir2 is part of a different protein complex, which does not include Sir3 or Sir4. The formation of silent chromatin leads to transcriptional repression at the silent mating types and at subtelomeric regions. The mating type of the yeast cell is determined by the mating type information (MATa or MATa) present at the mating locus (MAT) on chromosome III (12). The ability of a and a cells to respond to mating pheromones and mate depends on expression of the haploid-specific and MATa or MATa specific genes controlled by MATa or MATa transcription factors. In addition to being present at the active MAT locus, a copy of the MATa gene exists in a silent state at the silent HML locus and a copy of MATa exists at the silent HMR locus on the same chromosome (12). Normal diploid a/a state is defined by the coexpression of MATa and MATa information from the active MAT loci. However, the loss of silencing at the HMR and HML loci can create a pseudo diploid a/a state in a haploid cell by allowing the expression of MATa or MATa from the silent mating type loci. Accordingly, loss of silencing at the HML and HMR loci creates a mating defect and cells become insensitive to mating pheromones. The response of MATa cells to the alpha factor can be used as an assay for scoring derepressions at the silent HMR locus. Silencing at the rDNA locus has two important effects. It represses transcription of the ribosomal RNA gene or an inserted reporter gene (13), and it suppresses recombination between the tandem copies of rRNA genes (14). Beside the NAD-dependent deacetylase activity of Sir2, which is required for the silenced state at all silenced sites in the yeast

148

Newcomb and Bedalov

genome (15), other nuclear processes including DNA replication (16), nucleosome assembly (17) and remodeling machinery (18) and several other histone modifying enzymes also participate in silent chromatin formation and maintenance. Intriguingly, the main role for several histone modifying enzymes, such as Dot1 (19), Set1 (20, 21), and Sas2 (22–24), which are required for efficient silencing at telomeres, appears to be in limiting heterochromatin spreading. In their absence, the SIR complex extends beyond its normal boundaries and the redistribution of the SIR complexes weakens silencing at the native sites. While detailed mechanistic insight into how each of these and other enzymes contribute to the epigenetic memory is still lacking, it is clear that silencing of the reporter genes at telomeres is a very sensitive readout of the local chromatin state. 1.2. Silencing Assays in Yeast

Robust phenotypic assays have been developed for examining each of the silenced loci in the yeast genome (25). In the following sections we will discuss the assays that are suitable for high throughput small molecule screening. Table 1 lists the strains used to assay silencing at the telomeres, silent mating loci, and rDNA.

1.2.1. Silencing at the Telomeres

When a URA3 gene is introduced in the vicinity of a telomere its transcription is repressed (26). However, the transcriptional repression of a telomeric URA3 gene is overcome in synthetic medium lacking uracil (C-Ura). Low uracil growth conditions up regulate the transcription factor Ppr1, which transactivates URA3 and other genes required for uracil biosynthesis (28). Deletion of PPR1 can be used to modulate the strength of URA3 transcription. In the absence of PPR1, URA3 expression is largely dictated by the local chromatin state and is insensitive to the lack of uracil in cells. Accordingly, in the S288C strain background, the Dppr1

Table 1 Yeast strains used in silencing assays Assay

Relevant genotype

Selection criteria

Strain

Telomeric silencing

ppr1- DURA3-TEL (VII-L)

C -Ura media

UCC2210 (26)

URA3-TEL (VII-L)

C +5-FOA media

UCC3503 (26)

ADE2-TEL (V-R)

Colony color

UCC3503 (26)

HML silencing

Mat a

Sensitivity to a-factor

BY4741

HMR silencing

hmrD::TRP1/HMR

C -Trp media

YJB959 (27)

rDNA silencing

RDN1::Ty1-mURA3

C -Ura or C +5-FOA

S6 (13)

Identification of Inhibitors of Chromatin Modifying Enzymes

149

URA3-TEL (VII-L) reporter strain will grow poorly in media without uracil. In the Δppr1 strain, perturbed silencing at the telomeres permits sufficient URA3 transcription for growth in a medium lacking uracil. In addition to positive selection, the URA3 telomeric reporter strain can also be employed in a negative selection screen using 5-Flouroorotic Acid (5-FOA). When silencing is perturbed the URA3 gene products convert the nontoxic 5-FOA into 5-Flurouracil, which is highly toxic. Other telomeric reporter strains have been developed and include the ADE2-TEL (V-R) strain (26). The ADE2 telomeric reporter strain relies on a straightforward assay based on ADE2 expression and colony color. Strains expressing a wild type copy of ADE2 produce white colonies and strains with silenced or deleted ADE2 form red colonies. The ADE2-TEL (V-R) strain stochastically switches the ADE2 gene from being transcriptionally repressed to being transcriptionally active and this process leads to the variegated expression of the ADE2-TEL (V-R) gene. As the colonies grow, the transcriptional state of the telomere is passed on to daughter cells and red and white sectors appear in the colony (see Notes 1 and 2). The amount of sectoring in the colony is directly related to the transcriptional state of the ADE2 gene, and if silencing is perturbed colonies appear with less red sectoring (see Note 3). In addition to the URA3 and ADE2 reporter strains, other telomeric reporters, such as strains utilizing HIS3, or TRP1 reporters, are also available (reviewed in (25)) and can be easily adapted to screens for small molecules with antisilencing properties. 1.2.2. Silencing at HMR and HML Loci

Silencing at the HMR locus is assayed using a TRP1 reporter gene integrated at the HMR locus. In a cell with intact silencing, the TRP1 gene will not be expressed and the cell will be unable to grow in media lacking tryptophan (C –Trp). When silencing at the HMR locus is disrupted, the TRP1 gene will be expressed and the strain will grow in C –Trp media. Testing silencing at the HML locus relies on the mating pheromone a-factor. When two haploid yeast cells, of opposite mating type, come in close proximity to each other they secrete cell type specific mating pheromones (2). MATa cells secrete a-factor, a modified hydrophobic peptide, and MATa cells secrete a-factor, an unmodified soluble peptide. Typically a-factor is used in experiments utilizing a mating pheromone because it can be dissolved in liquid media and it is easily obtained commercially. a-Factor induces MATa cells to arrest in the G1 phase of the cell cycle and undergo morphological changes, known as shmooing, in preparation for cell and nuclear fusion. MATa cells with aberrant silencing at the HML locus are insensitive to a-factor and do not arrest in G1 or shmoo.

150

Newcomb and Bedalov

1.2.3. Silencing at the rDNA

Yeast ribosomal DNA (rDNA) consists of a 9.1 kb region of DNA, containing a 5S rRNA gene and a 35S pre-rRNA gene, which is repeated 100–200 times along chromosome XII. Silencing at the rDNA is assayed with a strain that utilizes a URA3 reporter, with a minimal TRP1 promoter (mURA3), integrated at the rDNA repeats (13). Use of the mURA3 instead of native URA3 improves URA3 silencing and permits scoring by positive and negative selection. In addition to silencing, recombination at the rDNA locus can be measured through the loss rate of an ADE2 gene integrated into the rDNA (29).

1.3. Screening Strategy

The use of a telomeric URA3 reporter as a primary screen has several advantages over other available reporter systems. Primarily, silencing at the telomere is a very sensitive readout of the chromatin state (see Note 4) that is influenced by the products of many genes. Furthermore, the loss of silencing-induced activation of the URA3 reporter can be scored as growth in medium lacking uracil, or as toxicity in medium containing 5-FOA. The design of a high throughput screen utilizing selection for growth has an advantage in that it requires that the compound inhibits a relevant target at concentrations that do not perturb other cellular processes (see Note 5).

1.3.1. Primary Screen

1.3.2. Secondary Assays

Compounds identified in the primary screen that promote growth in C –Ura, or demonstrate toxicity in 5-FOA are subjected to secondary screens with the goal of identifying compounds that specifically inhibit silencing. This is carried out by comparing toxicity of compounds in medium with and without 5-FOA, and by examining the ability of compounds to stimulate growth in medium lacking uracil. A dose response curve using the same concentration of drug employed in the primary screen and several twofold dilutions is carried out in complete medium, medium containing 5-FOA and medium lacking uracil. The concentration of an ideal antisilencing drug that inhibits growth by 50% (IC50) is expected to be at least 2–4-fold lower in medium containing 5-FOA relative to the IC50 in C-medium (Fig. 1). The dose response curve in medium lacking uracil is expected to identify the concentration of drug that activates the reporter and stimulates growth. Growth inhibition at the higher drug concentrations for compounds that activate a telomeric reporter may indicate inhibition of an essential activity that is important for telomeric silencing, or indicate inhibition of an unrelated essential target (off target activity). In order to demonstrate that a compound interferes specifically with the state of chromatin, and does not solely promote transactivation of the URA reporter, silencing of other reporters at telomeres and at loci other than the telomeres should be characterized. For this purpose, the sectoring assay using a strain containing ADE2 at the telomere can be used. The presence of

Identification of Inhibitors of Chromatin Modifying Enzymes

151

Fig. 1. (a) Example dose response curves for a compound identified in a primary screen. Curve 1 represents the dose response in C +5-FOA. The IC50 for the compound in C +5-FOA is 3.13 mM. Curve 3 represents the dose response in C media, expected if the compound used for Curve 1 is highly specific and nontoxic (e.g., highly specific Sir2 inhibitor with no off target activity). Curve 2 represents the dose response curve in C medium that would be expected if the compound used in Curve 1 had off target activity or activity against an essential enzyme involved with silencing (IC50 in C medium is 12.5). Fourfold greater sensitivity in C +5-FOA relative to C medium suggests activation of the URA3 reporter. (a) Curve 3 represents the dose response in C-Ura medium for a highly specific compound that inhibits silencing and that has no off target (e.g., highly specific Sir2 inhibitor with no off target activity). Curve 2 shows the expected dose response curve in C-Ura medium for a compound that at higher concentrations exhibits off target activity or that inhibits an enzyme that is, in addition to silencing, required for normal cell growth.

heterochromatin at sites other than the telomeres can be examined by monitoring activation of a TRP1 reporter integrated at the HMR locus or the responsiveness of MATa cells to a-factor can be used to assay silencing of the endogenous HML locus. Furthermore, both silencing and recombination rates at rDNA can be assayed using available strains. A URA3 reporter strain has been developed for assaying silencing, and strains have been developed with ADE2 integrated at the rDNA to determine the recombination rate by measuring the loss of the ADE2 gene. 1.4. Characterizing Hits 1.4.1. Identification of Molecular Targets

Identification of the targets of compounds identified through cell based screens presents one of the major challenges of chemical genetics. Traditional target identification strategies rely on biochemical affinity methods such as purification of the target protein using a drug affinity matrix. One of the major drawbacks of this approach is the need for ligand modification, which is often time consuming, and can lead to loss of drug activity. Furthermore, this approach is limited to compounds that have a relatively high affinity to their corresponding target. Beside the affinity purification methods, the large body of preexisting knowledge of yeast epigenetic regulation and the facile genetics offer additional means for target identification. The ideal drug, upon binding the target protein, mimics either a loss of function or a gain of function

152

Newcomb and Bedalov

mutation in the corresponding gene. This premise serves as a basis for identifying drug targets by drug-mutant matching approaches. An antisilencing drug that inhibits a specific chromatin modifying enzyme is expected to (1) recreate the effect of a point mutation in the corresponding enzyme that abrogates its enzymatic activity in vitro, (2) recreate the same global or local chromatin alterations, (3) change the global transcriptional profile, and (4) replicate the synthetic interaction profiles of their corresponding loss of function mutants. 1.4.2. In vitro Inhibition of Known Enzymatic Activities Important for Silencing

Several of the genes that affect telomeric silencing in yeast encode proteins that have defined in vitro enzymatic activities. While the published assays, such as enzymatic assays for Dot1, Set1, Sir2, and Sas2 described in references (19, 20, 23, 30, 31), may not be suitable for high throughput screening, the number of compounds that need to be characterized in this step is expected to be limited, which makes the in vitro testing for enzyme inhibition the most straightforward approach for evaluating whether the antisilencing compounds inhibit these known enzymatic activities.

1.4.3. Histone Modifications in Drug Treated Cells and Silencing Mutants

A deficiency of histone modifying enzymes is expected to lead to local (i.e., telomeric) or global alterations in specific histone posttranslational modifications (PTM). The ability of compounds to alter histone PTM can be evaluated using a panel of antibodies specific to different modifications of histone proteins using Western blots, for global, or chromatin immunoprecipitation for local telomere alterations respectively. Antibodies targeting more than 40 different histone modifications (methylation, acetylation of different lysine residues) are commercially available.

1.4.4. Matching of Transcriptional Effects

Silencing in yeast occurs at three locations: telomeres, silent mating loci and rDNA. While the loss of SIR2 affects silencing at all three sites, the effect of inactivating other chromatin modifying enzymes may be restricted to specific locations. In addition to known sites affected by silencing, proteins required for silencing also affect transcription, directly or indirectly, at other locations in the genome. As a result, the global transcriptional changes in drug treated wild type cells are expected to be highly correlated with the changes observed in cells with mutations in proteins required for transcription at various loci. The gene which, when deleted, creates a transcriptional profile most highly correlated with the transcription profile created by drug treated wild type cells should be further analyzed as potentially encoding for a drug target. Large numbers of yeast mutants have been subjected to transcriptional profiling (32, 33), and the data are readily available. The definite proof that the drug affects a putative target will rely on:

Identification of Inhibitors of Chromatin Modifying Enzymes

153

(a) Inhibition of the targets’ in vitro enzymatic activity, (b) Demonstrating specific binding of the drug to its target if no biochemical activity has been ascribed to the protein, (c) Obtaining a mutant protein through in vivo screening for alleles that confer drug resistance and have alterations in drug binding or are resistant to enzyme inhibition in vitro.

2. Materials 2.1. Positive Selection Screen

1. Strain UCC2210 Mat a ade2d::hisG his3d200 leu2d0 lys2d0 met15d0 trp1d63 ura3d0 adh4::URA3-TEL (VII-L) ppr1::HIS3 (see Note 6) 2. YEPD medium (For 2 L of media): 20 g bacto yeast extract (cat. no. DF0127-17-9, Fisher), 40 g bacto peptone, 40 g dextrose. 3. Complete synthetic (C) medium (for 2 L of media): 2.9 g yeast nitrogen base without amino acids and ammonium sulfate, 10 g ammonium sulfate, 40 g dextrose, and amino acids (omit Uracil for C –URA media) according to Table 2. 4. Flat bottom, clear 96-well microplate. 5. Compound library and a positive control, such as splitomicin (see Note 7), known to abrogate silencing in yeast. 6. 30°C incubator. 7. Microplate reader (e.g., Molecular Devices VERSAMAX Tunable Microplate Reader). 8. Distributing cells and compounds to 96-well plates is greatly simplified with the use of an automated 96-channel pipettor (e.g., a PlateMate or similar programmable device).

2.2. Negative Selection Screen

1. Strain UCC3503 Mat a ura3-52 lys2-801 ade2-101 leu2-d1 his3-d200 adh4::URA3-TEL (VII-L) ADE2-TEL (V-R) 2. C +5-FOA media is C media as described in Subheading 2.1, item 3 with 2 g 5-FOA (Zymoresearch) added to 2 L of C media. 3. See required materials for Positive Selection (Subheading 2.1).

2.3. Dose Response Assay

1. Strain UCC3503 Mat a ura3-52 lys2-801 ade2-101 leu2-d1 his3-d200 adh4::URA3-TEL (VII-L) ADE2-TEL (V-R) or Strain UCC2210 Mat a ade2d::hisG his3d200 leu2d0 lys2d0 met15d0 trp1d63 ura3d0 adh4::URA3-TEL (VII-L) ppr1::HIS3.

154

Newcomb and Bedalov

Table 2 Amino acids used for preparation of synthetic complete media Amino acid

Grams

Cat. No.

Vendor

Adenine (6-aminopurine)

1

A-8626

Sigma

L-arginine

1

BP370-100

Fisher

5

BP374-100

Fisher

5

BP378-100

Fisher

L-histidine

1

BP382-100

Fisher

L-isoleucine

4

BP384-100

Fisher

L-leucine

4

BP385-100

Fisher

L-lysine

3

L9037

Sigma

L-methionine

1

BP388-100

Fisher

L-phenylalanine

2.5

BP391-100

Fisher

L-serine

20

BP393-100

Fisher

L-threonine

10

BP394-100

Fisher

L-tryptophan

4

BP395-100

Fisher

L-tyrosine

3

BP396-100

Fisher

Uracil

1

U-1128

Sigma

L-valine

7.5

BP397-100

Fisher

L-aspartic

acid

L-glutamic

acid

2. C media, C –URA media, and C +5-FOA media as described above. 3. See required materials for Positive Selection (Subheading 2.1). 2.4. a-Factor Resistance Assay

1. Strain BY4741 Mat a his3 leu2 met15 ura3. 2. C media (described in Subheading 2.1). 3. a-Factor (Zymoresearch). 4. See required materials for Positive Selection (Subheading 2.1).

2.5. HMR Silencing Assay

1. Strain YJB959 MATa ura3-1 ade2-1 his3-11 leu 2-3,112 can 1-100 trp1-1 HMR::TRP1 2. C media prepared without tryptophan (see Subheading 2.1, item 3) 3. See required materials for Positive Selection (Subheading 2.1).

2.6. Red/White Colony Formation

1. Strain UCC3503 Mat a ura3-52 lys2-801 ade2-101 leu2- 1 his3- 200 adh4::URA3-TEL (VII-L) ADE2-TEL (V-R)

Identification of Inhibitors of Chromatin Modifying Enzymes

155

2. YEPD agar (For 2 L of media): 20 g bacto yeast extract, 40 g bacto peptone, 40 g dextrose, 40 g bacto agar. 3. YEPD agar plates containing drug at the IC50 concentration and plates without drug. 2.7. rDNA Silencing

1. Strain S26 Mat a his3 200 leu2 1 ura3-167 RDN1::Ty1mURA3. 2. C-Ura media, or +5-FOA media as described above in Subheading 2.2, item 2. 3. See required materials for Positive Selection (Subheading 2.1).

3. Methods 3.1. Primary Screens 3.1.1. Positive Selection Screen

1. Grow a culture of the assay strain in YEPD media to saturation (approximately 2 × 108 cells/mL). 2. Harvest cells by centrifugation and wash once with water. 3. Resuspend washed cell pellet in C-media prepared without uracil (referred to as C –URA) to a density of 5 × 104 cells/mL. 4. Deliver 135 mL of cell suspension to each well of a 96-well plate. 5. The compounds to be assayed should be diluted with water to 10× the desired assay concentration. Generally stock concentrations of compounds should be between 50 and 150 mM, solubility permitting. 6. Deliver 15 mL of the compound stocks to the yeast cells. 7. The total volume of each well will be 150 mL, with cells at a density of 5 × 104 and the compounds to be assayed at final concentrations in the range of 5–15 mM (see Note 7). 8. Incubate the plates at 30°C for 24–48 h (see Note 8). 9. Score growth by reading the optical density (OD) of the wells at 660 nm on a microplate reader (Molecular Devices VERSAMAX Tunable Microplate Reader) (see Notes 7 and 9).

3.1.2. Negative Selection Screen

Conducting a negative selection screen in 5-FOA media is carried out in the same fashion as the positive selection screen, but it is necessary to use a strain that is wild type for PPR1. Incubate plates for 24–48 h at 30°C (see Notes 7, 8, and 10).

3.2. Secondary Screens

1. Prepare cells as in the positive selection screen (Subheading 3.1.1), and resuspend in the appropriate media (see Note 11).

3.2.1. Dose Response Assay

2. Deliver 135 mL of cell suspension to each well of a 96-well plate.

156

Newcomb and Bedalov

3. Prepare 1:2 serial dilutions of the compound of interest, spanning a range of concentrations above and below the concentration used in the primary screen. 4. Deliver 15 mL of compound dilutions to cells (see Note 7). 5. Incubate at 30°C for 24–48 h and read the optical density (see Notes 9 and 10). 3.2.2. a-Factor Resistance (HML Silencing)

1. Prepare cells as in the positive selection screen (Subheading 3.1.1) and resuspend in C media containing 5 mM a-Factor. 2. Distribute 135 mL of cell suspension into each well of a 96-well plate and add 15 mL of the compounds of interest at the desired concentration (see Notes 7, 12, and 13). 3. Incubate the plate(s) at 30°C for 12–24 h, and read the optical density at 660 nm.

3.2.3. HMR Silencing Assay

1. Prepare cells as in the positive selection screen (Subheading 3.1.1) and resuspend in C media prepared without tryptophan. 2. Distribute 135 mL of cells to each well of a 96-well plate, and deliver 15 mL of the compounds of interest (see Note 7 and 12). 3. Incubate the plate(s) at 30°C for 12–36 h, and read the optical density at 660 nm.

3.2.4. Red/White Colony Formation Assay

The Red/White colony formation assay can be carried out using 96-well plate format. Grow cells in YEPD liquid to saturation. 1. Dilute and plate cells for singles (see Note 14) on YEPD agar and YEPD agar containing the drug of interest. 2. Grow colonies at 30°C for 24–48 h and transfer plates to 4°C for 72 h to develop red pigmentation (see Notes 1, 2, 3, and 7).

3.2.5. rDNA Silencing Assay

1. The S26 strain (Mat a his3 200 leu2 1 ura3-167 RDN1::Ty1mURA3) can be used in a positive selection screen in medium lacking uracil as described in Subheading 3.1.1, or in a negative selection screen in medium containing 5-FOA as described in Subheading 3.1.2.

4. Notes 1. Red-pigmented colonies do not grow as robustly as normal white colonies. The use of YEPD media instead of C media will help minimize the disparity. 2. The degree of Red/White sectoring is not uniform among colonies of the same strain.

Identification of Inhibitors of Chromatin Modifying Enzymes

157

3. Development of the red pigmentation is greatly improved by placing plates at 4°C for several days after colonies have formed. 4. We have noted that derepression of telomeric reporters can be achieved by at least fivefold lower concentrations of the Sir2 inhibitor splitomicin, than those required for derepression of the reporters at the silent mating loci. 5. The choice of a positive or negative selection strategy for the primary screen should be based on the characteristics of the chemical libraries being assayed. If the library is enriched for toxic compounds, (e.g., 20–25% of compounds in the National Cancer Institute repository exhibit toxicity), a primary screen using positive selection, which assays for stimulation of growth in medium lacking uracil, is expected to be more efficient than a screen for toxicity in 5-FOA medium. Libraries that are not enriched for toxic compounds can be efficiently screened for toxicity in 5-FOA as a primary screen followed by dose response evaluations of toxicity in C medium and growth enhancement in medium lacking uracil as secondary screens. 6. Deletion of PDR1, PDR3 which encode transcription factors that activate pleotrophic drug resistance genes, and ERG6 required for ergosterol biosynthesis may increase the intracellular drug concentrations by impairing drug efflux and enhancing membrane permeability. 7. The use of splitomicin is helpful to gauge the performance of the assay and to determine the stopping point of the assay. At early timepoints, such as 18–36 h, splitomicin will permit cells to grow in C-Ura media and untreated cells will exhibit no growth. At later timepoints (48 h and longer), the stochastic nature of silencing and the resulting spontaneous derepression of the URA3 reporter in a small fraction of cells will permit growth in some, though not all, negative control wells. Thus, scoring at later time points allows for greater sensitivity but may lead to an increased rate of false positive hits. In order to minimize false positives, growth should be scored at a timepoint when wells treated with splitomicin have an OD660 of 0.25-0.4 (on a Molecular Devices VERSAMAX Tunable Microplate Reader) and no growth is observed in control wells. Splitomicin can be used as a positive control at a concentration of 10 mM in Subheadings 3.1.1, 3.1.2, 3.2.2, 3.2.3, 3.2.4, and 3.2.5. For the dose response assays in Subheading 3.2.1, a range of splitomicin concentrations from 0.1 to 50 mM should be analyzed. Colonies growing on agar containing 10 mM splitomicin appear white. Splitomicin should be diluted in the media immediately prior to use as it rapidly hydrolyzes in aqueous solutions (half life 1–8 h depending on pH).

158

Newcomb and Bedalov

8. Place plates inside a clean plastic bag, or seal with parafilm to prevent evaporation during the incubation period. 9. In a positive selection screen, an OD between 0.25 and 0.4 indicates growth. 10. In a negative selection screen (growth in 5-FOA), >75% growth inhibition should be scored as a hit (growth inhibition = 1 − (ODtreated/ODuntreated). 11. For the dose response assays, resuspend cells in C media, C-Ura media and C +5-FOA media, and conduct the dose response assay in each type of media. 12. When assaying the compounds for silencing at the mating loci and ADE-TEL (V-R) locus, use the concentration of drug determined to best stimulate growth in C -Ura media in Subheading 3.2.1. When assaying compounds for rDNA silencing use the concentration from the dose response curve in C-Ura media that best stimulates growth, for a positive selection screen; or the IC50 concentration in 5-FOA, for a negative selection screen. 13. In addition to growth, loss of silencing is also expected to abrogate the ability of MATa cells to shmoo, which can be scored using a microscope. 14. For 96-well format, plate 5–10 cells per well.

Acknowledgments This work was supported by the National Institute of Health (CA129132-01A1) and the Leukamia and Lymphoma Society grants to A. Bedalov. References 1. Kouzarides, T. (2007) Chromatin modifications and their function Cell 128, 693–705. 2. Marsh, L., Neiman, A. M., and Herskowitz, I. (1991) Signal transduction during pheromone response in yeast Annual Review of Cell Biology 7, 699–728. 3. Schuettengruber, B., Chourrout, D., Vervoort, M., Leblanc, B., and Cavalli, G. (2007) Genome regulation by polycomb and trithorax proteins. Cell 128, 735–45. 4. Jones, P. A., and Baylin, S. B. (2007) The epigenomics of cancer Cell 128, 683–92. 5. Bedalov, A., Gatbonton, T., Irvine, W. P., Gottschling, D. E., and Simon, J. A. (2001)

Identification of a small molecule inhibitor of Sir2p. Proc Natl Acad Sci U S A 98, 15113–8. 6. Sauve, A. A., Wolberger, C., Schramm, V. L., and Boeke, J. D. (2006) The Biochemistry of Sirtuins. Annu Rev Biochem 75, 435–65. 7. Bedalov, A., Hirao, M., Posakony, J., Nelson, M., and Simon, J. (2003) NAD-dependent deacetylase Hst1p controls biosynthesis and cellular NAD levels in Saccharomyces cerevisiae. Mol Cell Biol 23. 8. Chang, C. R., Wu, C. S., Hom, Y., and Gartenberg, M. R. (2005) Targeting of cohesin by transcriptionally silent chromatin. Genes Dev 19, 3031–42.

Identification of Inhibitors of Chromatin Modifying Enzymes 9. Pruitt, K., Zinn, R. L., Ohm, J. E., McGarvey, K. M., Kang, S. H., Watkins, D. N., Herman, J. G., and Baylin, S. B. (2006) Inhibition of SIRT1 reactivates silenced cancer genes without loss of promoter DNA hypermethylation. PLoS Genet 2, e40. 10. Heltweg, B., Gatbonton, T., Schuler, A. D., Posakony, J., Li, H., Goehle, S., Kollipara, R., DePinho, R. A., Gu, Y., Simon, J. A., and Bedalov, A. (2006) Antitumor activity of a small molecule inhibitor of human Sir2 enzymes. Cancer Res 66, 4368–77. 11. Rusche, L. N., Kirchmaier, A. L., and Rine, J. (2003) The establishment, inheritance, and function of silenced chromatin in Saccharomyces cerevisiae. Annu Rev Biochem 72, 481–516. 12. Herskowitz, I. (1988) Life cycle of the budding yeast Saccharomyces cerevisiae. Microbiol Rev 52, 536–53. 13. Smith, J. S., and Boeke, J. D. (1997) An unusual form of transcriptional silencing in yeast ribosomal DNA. Genes Dev 11, 241–54. 14. Gottlieb, S., and Esposito, R. E. (1989) A new role for a yeast transcriptional silencer gene, SIR2, in regulation of recombination in ribosomal DNA. Cell 56, 771–6. 15. Imai, S., Armstrong, C. M., Kaeberlein, M., and Guarente, L. (2000) Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase. Nature 403, 795–800. 16. Suter,B.,Tong,A.,Chang,M.,Yu,L.,Brown,G.W., Boone, C., and Rine, J. (2004) The origin recognition complex links replication, sister chromatid cohesion and transcriptional silencing in Saccharomyces cerevisiae. Genetics 167, 579–91. 17. Sharp, J. A., Fouts, E. T., Krawitz, D. C., and Kaufman, P. D. (2001) Yeast histone deposition protein Asf1p requires Hir proteins and PCNA for heterochromatic silencing. Curr Biol 11, 463–73. 18. Dror, V., and Winston, F. (2004) The Swi/ Snf chromatin remodeling complex is required for ribosomal DNA and telomeric silencing in Saccharomyces cerevisiae. Mol Cell Biol 24, 8227–35. 19. van Leeuwen, F., Gafken, P. R., and Gottschling, D. E. (2002) Dot1p modulates silencing in yeast by methylation of the nucleosome core. Cell 109, 745–56. 20. Krogan, N. J., Dover, J., Khorrami, S., Greenblatt, J. F., Schneider, J., Johnston, M., and Shilatifard, A. (2002) COMPASS, a histone H3 (Lysine 4) methyltransferase required for telomeric silencing of gene expression. J Biol Chem 277, 10753–5. 21. Venkatasubrahmanyam, S., Hwang, W. W., Meneghini, M. D., Tong, A. H., and Madhani,

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

159

H. D. (2007) Genome-wide, as opposed to local, antisilencing is mediated redundantly by the euchromatic factors Set1 and H2A.Z. Proc Natl Acad Sci U S A 104, 16609–14. Reifsnyder, C., Lowell, J., Clarke, A., and Pillus, L. (1996) Yeast SAS silencing genes and human genes associated with AML and HIV-1 Tat interactions are homologous with acetyltransferases. Nat Genet 14, 42–9. Sutton, A., Shia, W. J., Band, D., Kaufman, P. D., Osada, S., Workman, J. L., and Sternglanz, R. (2003) Sas4 and Sas5 are required for the histone acetyltransferase activity of Sas2 in the SAS complex. J Biol Chem 278, 16887–92. Kimura, A., Umehara, T., and Horikoshi, M. (2002) Chromosomal gradient of histone acetylation established by Sas2p and Sir2p functions as a shield against gene silencing. Nat Genet 32, 370–7. van Leeuwen, F., and Gottschling, D. E. (2002) Assays for gene silencing in yeast. Methods Enzymol 350, 165–86. Singer, M. S., Kahana, A., Wolf, A. J., Meisinger, L. L., Peterson, S. E., Goggin, C., Mahowald, M., and Gottschling, D. E. (1998) Identification of high-copy disruptors of telomeric silencing in Saccharomyces cerevisiae. Genetics 150, 613–32. Buck, S. W., and Shore, D. (1995) Action of a RAP1 carboxy-terminal silencing domain reveals an underlying competition between HMR and telomeres in yeast. Genes Dev 9, 370–84. Aparicio, O. M., and Gottschling, D. E. (1994) Overcoming telomeric silencing: a trans-activator competes to establish gene expression in a cell cycle-dependent way. Genes Dev 8, 1133–46. Kaeberlein, M., McVey, M., and Guarente, L. (1999) The SIR2/3/4 complex and SIR2 alone promote longevity in Saccharomyces cerevisiae by two different mechanisms. Genes Dev 13, 2570–80. Smith, J. S., Avalos, J., Celic, I., Muhammad, S., Wolberger, C., and Boeke, J. D. (2002) SIR2 family of NAD(+)-dependent protein deacetylases. Methods Enzymol 353, 282–300. Alcasabas, A. A., Osborn, A. J., Bachant, J., Hu, F., Werler, P. J., Bousset, K., Furuya, K., Diffley, J. F., Carr, A. M., and Elledge, S. J. (2001) Mrc1 transduces signals of DNA replication stress to activate Rad53. Nat Cell Biol 3, 958–65. Mnaimneh, S., Davierwala, A. P., Haynes, J., Moffat, J., Peng, W. T., Zhang, W., Yang, X., Pootoolal, J., Chua, G., Lopez, A., Trochesset, M., Morse, D., Krogan, N. J., Hiley, S. L., Li, Z., Morris, Q., Grigull, J., Mitsakakis, N., Roberts, C. J., Greenblatt, J. F., Boone, C.,

160

Newcomb and Bedalov

Kaiser, C. A., Andrews, B. J., and Hughes, T. R. (2004) Exploration of essential gene functions via titratable promoter alleles. Cell 118, 31–44. 33. Hughes, T. R., Marton, M. J., Jones, A. R., Roberts, C. J., Stoughton, R., Armour, C. D., Bennett, H. A., Coffey, E., Dai, H., He, Y. D.,

Kidd, M. J., King, A. M., Meyer, M. R., Slade, D., Lum, P. Y., Stepaniants, S. B., Shoemaker, D. D., Gachotte, D., Chakraburtty, K., Simon, J., Bard, M., and Friend, S. H. (2000) Functional discovery via a compendium of expression profiles. Cell 102, 109–26.

Chapter 9 Exploiting Yeast Genetics to Inform Therapeutic Strategies for Huntington’s Disease Flaviano Giorgini and Paul J. Muchowski Summary Huntington’s disease (HD) is a devastating neurodegenerative disorder that is inherited in an autosomal dominant fashion and is caused by a polyglutamine expansion in the protein huntingtin (htt). In recent years, modeling of various aspects of HD in the yeast Saccharomyces cerevisiae has provided insight into the conserved mechanisms of mutant htt toxicity in eukaryotic cells. The high degree of conservation of cellular and molecular processes between yeast and mammalian cells have made it a valuable system for studying basic mechanisms underlying human disease. Yeast models of HD recapitulate conserved disease-relevant phenotypes and can be used for drug discovery efforts as well as to gain mechanistic and genetic insights into candidate drugs. Here we provide a detailed overview of yeast models of mutant htt misfolding and toxicity and the molecular and phenotypic characterization of these models. We also review how these models identified novel therapeutic targets and compounds for HD and discuss the benefits and limitations of this model genetic system. Finally, we discuss how yeast may be used to provide further insight into the molecular and cellular mechanisms underlying HD and treatment strategies for this devastating disorder. Key words: Huntington’s disease, Huntingtin, Yeast models, Protein misfolding, Neurodegeneration

1. Introduction Although the Huntington’s disease (HD) gene was cloned over a decade ago (1), no therapeutic options are available for HD patients. HD is caused by an expansion of a polyglutamine (polyQ) stretch in the protein huntingtin (htt), which leads to misfolding and aggregation of mutant htt. Many studies support the hypothesis that mutant htt causes HD through a gain-offunction mechanism. Models of HD have been developed by

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_9, © Humana Press, a part of Springer Science + Business Media, LLC 2009

161

162

Giorgini and Muchowski

expressing mutant htt or mutant htt fragments in mice, mammalian cells, Drosophila, Caenorhabditis elegans, and the baker’s yeast Saccharomyces cerevisiae. Yeast has been used for centuries to produce foods such as bread and wine (2). Over the past 40 years, studies in yeast have provided critical insight into basic cellular processes that are conserved in higher eukaryotes, including cell division, replication, metabolism, protein folding, and intracellular transport (3). The conserved cellular mechanisms elucidated in these organisms, such as cell-cycle regulation, have played a major role in understanding cancer and other complex diseases (4, 5). More recently, yeast has been used to model various aspects of protein misfolding disorders, such as HD and Parkinson’s disease (PD), and these models have been used for high-throughput screens to identify candidate therapeutic targets and candidate drugs. Yeast is an ideal model organism for biological research, owing to its power and ease of use for genetic studies. Classical genetics is easily performed in S. cerevisiae because it can exist in both haploid and diploid states and because it is easy to mate haploid strains and sporulate diploid strains. Yeast is perhaps the best-characterized eukaryotic organism. Its ~14 Mb genome was the first eukaryotic genome to be fully sequenced (6), and Saccharomyces genes and proteins are extensively annotated in several genomic and proteomic databases. Other useful properties of yeast include rapid growth on defined media, the ease of replica plating and mutant isolation, and importantly, a highly versatile DNA transformation system (7). In addition, genes of interest can easily be knocked out or altered by homologous recombination (7). These features, along with the ease of mutagenesis and the availability of several mutant collections and open reading frame (ORF) libraries, allow for rapid isolation of genetic modifiers of specific processes or phenotypes. In addition, molecular genetic manipulations, such as DNA transformation, targeted disruption of specific genes, generation of point mutations in cloned genes, and overexpression of proteins of interest can be performed in a matter of days, as compared to months or years in other model organisms. Development of arrayed libraries for systematic deletion or overexpression of most yeast genes has permitted rapid genomic screening and identification of novel gene interactions and drug targets. Here we review how studies in yeast have allowed dissection of the basic underlying mechanisms of mutant htt toxicity and have aided in the identification of novel small molecules and drug targets with therapeutic potential for HD.

Exploiting Yeast Genetics to Inform Therapeutic Strategies

163

2. Huntington’s Disease HD is a fatal neurodegenerative disorder that predominantly affects brain cells of the striatum and cortex, resulting in progressive chorea, rigidity, and dementia (8). The prevalence of HD is 5–10 in 100,000 people in the United States and United Kingdom (9). In adults, the median age of onset is the late forties or early fifties, but the juvenile form can appear as early as 2 years of age (10). The mutation that causes HD is an expansion of CAG repeats in the HD gene and is inherited in an autosomal dominant manner (1). In the general population, the number of CAG repeats ranges from 4 to 35. Expansions of 36–39 CAG repeats increase the risk of developing HD, while expansions of 40 or more CAG repeats are fully penetrant (11, 12). The length of the CAG repeat correlates inversely with the age of onset (13), with expansions of 70 CAG repeats or longer inevitably leading to juvenile-onset HD (1, 14). Although it is inversely correlated with the number of CAG repeats, the age of onset is highly variable when controlling for repeat length. Indeed, work with HD kindreds containing over 18,000 individuals found that approximately 40% of variation in age of onset in HD patients is due to genetic modifiers (15), suggesting that many therapeutic targets may be available. Pathologically, HD is characterized by a 10–20% loss of brain mass. The atrophy occurs in many brain regions but is most prominent in the striatum (16). Medium spiny neurons, which account for about 95% of striatal cells, are the most affected cells. These cells, which express the g-aminobutyric acid (GABA) receptor, are inhibitory projection neurons that carry the output of the striatum to the globus pallidus and the substantia nigra (16). The striatal pathology probably underlies the involuntary movements and several other symptoms of the disease. The major neuropathological hallmark in HD is brain lesions composed of intranuclear and cytoplasmic inclusions that contain mutant htt. The CAG repeat in the HD gene encodes a polyQ stretch at the N-terminal end of htt. A protein fragment encoding the first exon of htt with an expanded polyQ can spontaneously form amyloid fibers in vitro (17), and aggregates containing mutant htt are present in the brains of mice expressing mutant htt (18) and in brains of HD patients (19). The length of the polyQ expansion in htt correlates directly with its aggregation kinetics in vitro and with severity of the disease. The 17-amino acid amino-terminus of htt, which lies directly upstream of the polyglutamine tract, forms an amphipathic a-helical domain that can bind membranes and be targeted

164

Giorgini and Muchowski

to vesicles and the endoplasmic reticulum (20, 21). This domain appears to modulate the toxicity and aggregation of mutant htt, and it disrupts calcium homeostasis in glutamate-challenged PC12 cells (20). Immediately after the polyglutamine tract, htt contains two proline-rich regions; in mutant htt, these regions have been implicated in aggregation, cellular toxicity, sequestration of trafficking proteins, and membrane interactions (22–28). Further downstream of the proline-rich region, htt contains several HEAT domains, each approximately 40 amino acids long, that form hydrophobic a-helices and are likely to be involved in protein interactions (29). The structure of htt led to the prediction that wild-type htt protein is a multifunctional scaffolding protein (30, 31). Despite the many putative roles of htt in the cell, the pathogenesis of HD appears to reflect predominantly a gain-of-function mechanism due to expansion of the polyQ tract rather than a loss of wild-type htt function (32). Several molecular mechanisms have been implicated in the pathogenesis of HD, including transcriptional dysregulation, perturbations of kynurenine pathway metabolites which result in excitotoxicity, oxidative stress, impaired energy metabolism, defective vesicle trafficking in axons, and impairment of ubiquitination and proteasomal function.

3. Using Yeast to Model Neurodegenerative Diseases

Yeast have been used to model features of several neurodegenerative disorders, including HD, PD, Alzheimer’s disease (AD), Friedreich’s ataxia, and amyotrophic lateral sclerosis (33, 34). Additionally, studies in yeast using both yeast prions and mammalian prion-related protein have provided insight into the mechanism of prion action and disease (35, 36). While models of disease in higher eukaryotes (fruit flies, nematodes, and mice) are extremely powerful tools for analysis of genetic and pharmacological modification of phenotype/pathology, yeast models provide a rapid and facile alternative to the higher cost and slower pace of these models. Yeast models of protein misfolding allow both genetic and pharmacological screens with a rapidity and ease not possible in fly and worm models, and can be used to identify conserved cellular mechanisms critical to toxicity and candidate therapeutic targets (37–39). Ultimately, candidate modifier genes or compounds identified by screens in yeast must be validated in mammalian systems. Aspects of neurodegenerative diseases have been modeled in yeast by analyzing the function of yeast homologs of human disease genes or by analyzing the phenotypes caused by expressing

Exploiting Yeast Genetics to Inform Therapeutic Strategies

165

human disease genes in yeast. The first approach has been used to model aspects of Friedreich’s ataxia, amyotrophic lateral sclerosis, and prion disease. Studies of the yeast homologs of the human genes implicated in Friedreich’s ataxia (YHF1) and amyotrophic lateral sclerosis (SOD1) have helped elucidate pathogenic mechanisms underlying those diseases (33, 40). In the second approach, genes of interest are expressed in yeast, and this has proven successful for modeling aspects of HD, PD, and AD. These models recapitulate many disease-relevant phenotypes and have provided mechanistic insight into the pathology of these diseases. Obviously, there are limitations to studying mechanisms of protein misfolding in yeast, as there are with any model system. First, genes involved in neurodegeneration may not be present in the yeast genome. Nonetheless, several promising therapeutic targets and compounds relevant to HD have been identified in yeast, indicating that targeting evolutionarily conserved genes and pathways in yeast is a valid approach for drug discovery. Second, cellular toxicity in yeast may not be related to that involved in neurodegeneration. However, yeast can undergo apoptosis-like cell death in response to several stimuli, and several yeast orthologs of crucial apoptotic regulators exist (41). Intriguingly, expression of a toxic mutant htt fragment in yeast produces hallmarks of apoptosis, including mitochondrial fragmentation and caspase activation (42). In addition, expression of wild-type a-synuclein or the inherited mutants (A30P, A53T) in a yeast model of PD triggers several markers of apoptosis, and deletion of a yeast metacaspase gene suppresses many of these apoptosis-like phenotypes (43). In summary, several thorough and insightful studies have shown that yeast can be used to model pathogenic mechanisms involved in neurodegenerative diseases.

4. Overview of Yeast Models of HD and Phenotypic Analyses

Several yeast models of HD have been developed to study the folding and behavior of mutant htt within a eukaryotic cell (44–47) (Table 1).These studies have supported and provided further insight into observations made with mutant htt in mammalian cells. Expression of an amino-terminal fragment of htt containing the polyQ region fused to green fluorescent protein (GFP) in yeast results in a polyQ length–dependent aggregation and formation of cytoplasmic and nuclear inclusions that can easily be detected in living cells by fluorescence microscopy (42, 44, 46, 47). Genetic impairment of the ubiquitin/proteasome degradation pathway via lesions in the genes encoding the ubiquitin/ proteasome pathway components UBA1, DOA3, and SEN3 does

166

Giorgini and Muchowski

Table 1 Studies in which yeast were used to model aspects of HD Model

IBsa

Aggregatesb

Localization

Toxicity

Krobitsch et al. (2000)

+

+

Cytoplasmic



Muchowski et al. (2000)

+

+

Cytoplasmic



Hughes et al. (2001)

+

n/a

Cytoplasmic/nuclear c



Meriin et al. (2002)

+

+

Cytoplasmic/nuclear d

+

IBs inclusion bodies “+” indicates polyglutamine length-dependent formation of IBs based upon microscopy b “+” indicates polyglutamine length-dependent aggregation confirmed by biochemical approaches c SV40 nuclear localization signal used to target htt constructs to the nucleus d Sokolov et al., 2006 a

not alter the formation of mutant htt inclusions function in yeast, suggesting that ubiquitination is not required for their formation (44). A simple filter-retention assay has also been used to isolate detergent-insoluble aggregates formed by polyQ proteins with expanded repeats in the disease-causing range (45). In several of these models, modulation of chaperone expression alters the formation of mutant htt aggregates or inclusions. Co-expression of Hsp40 (Ydj1) or Hsp70 (Ssa1 or Ssb1) with mutant htt (53Q) in yeast inhibited formation of these detergent-insoluble aggregates and resulted in the formation of detergent-soluble aggregates (45). In addition, Hsp40 and Hsp70 family members co-immunoprecipitated with the mutant htt fragment but not with the wild-type control. In another study, co-expression of Hsp40 and Hsp70 family members (Sis1 and Ssa1, respectively) with mutant htt fragments containing longer polyQ expansions (72Q and 103Q) resulted in a larger number of smaller inclusions (44). Overexpression of the chaperone Hsp104 increased the number of inclusions containing mutant htt, while deletion of Hsp104 eliminated their formation and resulted in diffuse GFP fluorescence, most likely due to the indirect effect of curing the yeast prion [RNQ+] (see below). Depletion of the chaperonin TRiC increases aggregation of a mutant htt fragment in yeast (48). Therefore, as in mammalian cells, expression of mutant htt fragments in yeast results in polyQ-length dependent aggregation that can be modulated by chaperone expression. Perhaps the most-studied and best-characterized yeast model of HD, the Meriin model (46), exhibits polyQ length–dependent toxicity. Expression of an htt fragment with a polyQ length in the pathogenic range (Htt103Q) is toxic in yeast, while expression

Exploiting Yeast Genetics to Inform Therapeutic Strategies

167

of the same construct containing a nonexpanded polyQ tract (Htt25Q) shows no negative effect on growth (46). This model has allowed genetic characterization of conserved cellular pathways required for mutant htt toxicity in yeast. Aggregation of mutant htt (103Q) was suppressed when the activities of certain chaperone proteins were impaired, including Hsp104, Sis1, and Ydj1, supporting observations made in the models discussed above. Interestingly, the polyQ length–dependent toxicity in the Meriin model requires the presence of the yeast prion Rnq1 in its prion conformation, as well as other putative prion proteins (39, 46). This model is unique among the yeast models in that it exhibits polyQ length-dependent toxicity, and we have focused much effort on understanding how protein misfolding/aggregation in this model leads to cellular toxicity. We learned that both cis and trans factors are critical for mutant htt toxicity in yeast. First, we found that several aggregation prone proteins play a role in mutant htt toxicity and affect the modulation of aggregation (27, 39). Several of these proteins are known or predicted yeast prion proteins. We and others have also observed that the flanking amino acid sequences are required for modulating toxic conformations of these htt fragments in yeast (25, 27). This conversion of benign species to toxic species (and vice-versa) can operate in cis or in trans (27). Several factors downstream of mutant htt misfolding/aggregation may contribute to toxicity in this yeast model of HD, including perturbations in the kynurenine pathway (39), defects in endocytosis (49, 50), transcriptional dysregulation (47), active transport along microtubules (51), apoptotic-like events (42), increased levels of reactive oxygen species (ROS) (39, 52), and mitochondrial dysfunction (42, 52). Our work dissecting the central role of the kynurenine pathway in mutant htt toxicity in yeast is described at length below. Using the Meriin model, we performed a genomic screen for null mutations that suppress the toxicity of Htt103Q (39). One of the most potent suppressors was a deletion of the gene BNA4, which encodes the yeast homolog of kynurenine 3-monooxygenase (KMO), which functions in the kynurenine pathway of tryptophan degradation. Intriguingly, this pathway is activated in HD patients and in animal models of HD (53). Several tryptophan metabolites are neuroactive, including 3-hydroxykynurenine (3-HK) and quinolinic acid (QUIN), whose neurotoxic effects are mediated by mechanisms that are likely to include the generation of reactive oxygen species (ROS). Kynurenine pathway metabolites and enzymes are well conserved from yeast to humans, and the genetics of the pathway have been extensively characterized in yeast (54). Since deletion of KMO is predicted to eliminate the synthesis of QUIN and 3-HK, the identification of bna4 as a suppressor

168

Giorgini and Muchowski

led us to hypothesize that upregulation of the QUIN branch of the kynurenine pathway could contribute to toxicity in yeast, as observed in HD patients and in mouse models of HD. If so, increased levels of QUIN and 3-HK could contribute to Htt103Q-dependent toxicity, and elimination of these metabolites by deletion of KMO activity could be suppress the toxicity. To test this hypothesis, we measured 3-HK and QUIN levels in wild-type yeast expressing Htt25Q or Htt103Q and in the bna4 suppressor strain expressing Htt103Q. Excitingly, 3-HK and QUIN levels were significantly higher in wild-type cells expressing the toxic Htt103Q construct than in controls expressing nontoxic Htt25Q, as in HD patients and in the mouse models of HD. As predicted, deletion of BNA4 eliminated 3-HK and QUIN and reduced Htt103Q-mediated toxicity. This was the first evidence for a direct relationship between lower levels of endogenous 3-HK and QUIN and reduced toxicity of mutant htt-dependent toxicity in a genetic model of HD. We also measured ROS levels using two oxidation-sensitive dyes: dihydroethidium and dihydrorhodamine-123. ROS levels were approximately eightfold higher in wild-type yeast expressing Htt103Q, than in Htt25Q controls; however, the levels of ROS in bna4 D yeast expressing Htt103Q were not significantly elevated. Thus, there is a direct correlation between the levels of 3-HK and QUIN, the levels of ROS, and the levels of cellular toxicity in this yeast model of HD. Interestingly, KMO is an outer mitochondrial membrane protein and mitochondrial dysfunction has been implicated in HD pathology and the generation of ROS. These findings suggest that expression of Htt103Q leads to upregulation of the kynurenine pathway, causing cellular toxicity through a ROSmediated mechanism. We also found that pharmacological inhibition of KMO in yeast with the compound Ro 61-8048 decreased levels of 3-HK, decreased generation of ROS, and ameliorated the Htt103Q growth impairment. Ro 61-8048 is currently in trials with HD model mice, and preliminary results indicate that though this compound does not readily cross the blood brain barrier, it improves several outcome measures in these mice (unpublished observations). Many of the above observations made in yeast have since been validated in other models of HD, including flies, mammalian cells, and mice, highlighting the value of yeast as a tool for studying conserved pathogenic mechanisms in HD and human disorders in general. Such confirmation suggests that these models can be reliably used to screen for both genes and compounds that modulate the toxicity and aggregation of mutant htt.

Exploiting Yeast Genetics to Inform Therapeutic Strategies

5. Using Yeast Models of HD for Genomic Screens and for Identifying Therapeutic Compounds

169

Yeast models of HD have enabled us to perform genomic screens to identify gene deletions that either enhance or suppress mutant htt toxicity in yeast (37, 39) (Fig. 1). These models have also allowed for screening and individual testing of small molecules to identify novel candidate therapeutic compounds which have been subsequently validated in Drosophila and mammalian cells (38, 55). The genetic screens have taken advantage of a collection of yeast gene deletion strains (YGDS) developed by the Saccharomyces Genome Deletion Project (56, 57). The YGDS collection contains 4,850 viable mutant haploid strains, each lacking a single gene, and has been used to identify new genes and pathways in several biological processes. To identify gene deletions that enhance mutant htt toxicity, we used a nontoxic exon 1 mutant htt fragment construct to screen the YGDS collection (37) (Fig. 1). This enhancer screen identified gene deletions that produced polyQ length-dependent

WT Strain + mHtt

mHtt

KO Strain + mHtt

Gene X

mHtt

mHtt

Not induced

mHtt is toxic No Growth

mHtt is not toxic Induced

Gene X Δ

Growth

Suppressor Screen

Growth

Gene X Δ

Induced

No Growth

Enhancer Screen

Fig. 1. Overview of loss-of-function screens for suppressors or enhancers of toxicity of a mutant htt fragment (mHtt) in yeast. A plasmid containing mHtt under the control of an inducible promoter is transformed into wild-type yeast cells and the YKO deletion set. If mHtt is toxic to wild-type yeast cells, then the YKO set can be screened for gene deletions (i.e., Gene X) that restore growth or suppress toxicity. If mHtt is not toxic to wild-type cells, the YKO set can be screened for gene deletions that enhance the toxicity of mHtt.

170

Giorgini and Muchowski

synthetic lethality in combination with the mutant htt fragment. The YGDS strains were transformed with the mutant htt construct, and 52 gene deletions that enhanced toxicity of the mutant htt fragment were isolated. Forty (77%) of these enhancers are genes with known or predicted functions; 35% belong to the functionally related categories of protein folding, response to cell stress, and ubiquitin-dependent protein catabolism; the rest were dispersed among several diverse functional categories. Among the suppressors were two Hsp40 homologs (Apj1 and Hlj1), suggesting that Hsp40 chaperones are necessary for suppression of polyQ toxicity. Interestingly, Hsp40 chaperones cooperate with Hsp70 partners, which modulate pathogenesis in mouse models of polyglutamine disorders (58, 59). We also screened the YGDS for loss-of-function suppressors of toxicity in the Meriin model (39, 46) (Fig. 1). The primary goal was to identify targets that, when genetically inhibited (deleted), reduce the toxicity of a mutant htt fragment and would presumably produce a similar phenotype if inhibited pharmacologically. We also sought to identify cellular pathways required for this toxicity. Of the genes identified, 25% encode proteins that cluster into the functionally related categories of vesicular transport, vacuolar protein sorting, and vacuolar import; 25% encode proteins involved directly in transcription or in the establishment/ maintenance of chromatin architecture; and ~21% encode known yeast prions or proteins containing Q/N-rich regions that may mediate prion-like aggregation. The genes encoding proteins involved in vesicular transport and vacuolar function are of particular interest because induction of autophagy in fly and mouse models of HD reduces the toxicity of expanded polyglutamine (60). Thus, in yeast, Htt103Q might be targeted for degradation in the vacuole, and the autophagic machinery might become saturated or impaired as Htt103Q accumulates in cells. We hypothesize that perturbation of protein sorting to the vacuole, a phenotype found in many of these gene deletion strains, may allow for resumption of normal autophagy, perhaps by excluding Htt103Q from the vacuole. Consistent with this model, deletion of the gene encoding the vacuolar protease Cps1 enhances the toxicity of a mutant htt fragment (37). For both the enhancer and suppressor screens described above, over half of the suppressors have mammalian homologs (61) In the case of the suppressors, ~87% are known to be expressed in the brain. Recently, a high-throughput screen of 16,000 compounds was performed to identify small-molecule inhibitors of polyQ aggregation in yeast expressing a GFP-tagged mutant htt fragment (38). This screen yielded nine small molecules that ameliorated growth and/or fluorescence. Upon further testing, one of these compounds (C2-8) suppressed polyQ aggregation in a PC12 cell–based

Exploiting Yeast Genetics to Inform Therapeutic Strategies

171

model and in neurons of cultured brain slices from HD model mice; it also ameliorated neurodegeneration in a Drosophila model of HD. Intriguingly, C2-8 is a structural analog of the aforementioned KMO inhibitor that ameliorates Htt103Q toxicity. This finding suggests that genetic and chemical-genetic approaches can converge at the same candidate targets. Moreover, C2-8 readily crosses the blood-brain barrier in mice and improves both motor performance and neuronal atrophy in a mouse model of HD (62). As predicted by the aggregation assay used to identify C2-8, the volume of aggregate in striatal neurons was reduced. A yeast model has also been used to validate green tea (−)-epigallocatechin-gallate (EGCG) as a therapeutic compound for HD (55). A library of ~5,000 natural substances was screened with a membrane filer retardation assay. One purified natural compound (EGCG) inhibited aggregation of mutant htt exon 1. When tested in yeast expressing a toxic mutant htt fragment (Htt72Q), EGCG improved growth and reduced the number of cells carrying Htt72Q inclusion bodies. Further tests showed that EGCG reduces neurodegeneration and improves motor impairment in a fly model of HD. These studies highlight the potential for yeastbased approaches to screen for candidate therapeutic compounds for treatment of neurodegenerative disorders. In summary, yeast models of HD have been very useful for identifying genetic modifiers of mutant htt toxicity and compounds that modulate the toxicity and aggregation of mutant htt. This work has led to further validation studies of KMO inhibitors, EGCG, and C2-8 as potential therapies for HD. Thus, despite their simplicity, yeast models are an important tool for identifying and characterizing therapeutics targets and compounds for HD.

6. Conclusions and Future Directions Within the last decade, several models in S. cerevisiae have been developed that recapitulate important features characterized by HD. The success of genetic and compound screens will likely lead to further exploitation of this model in the hopes of informing HD therapeutics in the future. A number of genetic approaches can be implemented with these models that have not yet been attempted, including overexpression screens with ORF libraries. Random clone libraries have been used for overexpression studies in yeast, but only recently has the Yeast ORF Collection (Open Biosystems) become available, an array of ~5,500 overexpression constructs covering the vast majority of the yeast genome. Systematic screening of the collection is efficient as the clones are arrayed in 96-well microtiter plates. Isolation of overexpression suppressors

172

Giorgini and Muchowski

of mutant htt toxicity would identify candidate ORFs that, when overexpressed, protect against cellular toxicity in yeast. So far, only nonessential genes have been tested in loss-offunction screens. Therefore many important targets have not been screened, and approaches need to be implemented to identify essential modifier genes in yeast. Critically, candidate modifiers or compounds identified in yeast need to be validated in more physiologically relevant models of neurodegeneration, such as mammalian cell and mouse models of HD. Nonetheless, yeast will undoubtedly continue to be useful for advancing our knowledge of the underlying cellular mechanisms in HD and many other human diseases, even in cases in which it is initially difficult to conceive of the utility of this simple, yet powerful organism. The promising candidate targets and compounds for HD identified in yeast thus far have provided further support for the role of yeast as “living test tubes” in studies of human disease.

References 1. A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington’s disease chromosomes. The Huntington’s Disease Collaborative Research Group. Cell 1993;72(6):971–83. 2. Botstein D, Fink GR. Yeast: an experimental organism for modern biology. Science 1988; 240(4858):1439–43. 3. Fields S, Johnston M. Cell biology. Whither model organism research. Science 2005; 307(5717):1885–6. 4. Hartwell LH. Nobel Lecture. Yeast and cancer. Biosci Rep 2002;22(3–4):373–94. 5. Nurse PM. Nobel Lecture. Cyclin dependent kinases and cell cycle control. Biosci Rep 2002;22(5–6):487–99. 6. Goffeau A, Barrell BG, Bussey H, et al. Life with 6000 genes. Science 1996;274(5287) :546,63–7. 7. Sherman F. Getting started with yeast. In: Guthrie CaF, G.R. (Eds), Methods Enzymol 1991;194:3–21. 8. Martin JB, Gusella JF. Huntington’s disease. Pathogenesis and management. N Engl J Med 1986;315(20):1267–76. 9. Harper PS. The epidemiology of Huntington’s disease. In: Bates G, Harper PS, Jones L, eds. Huntington’s Disease. Oxford, UK: Oxford University Press; 2002:159–97. 10. Kremer B. Clinical neurology of Huntington’s disease. In: Bates G, Harper PS, Jones L,

11.

12.

13.

14.

15.

16.

eds. Huntington’s disease. Oxford: Oxford University Press; 2002:28–61. Rubinsztein DC, Leggo J, Coles R, et al. Phenotypic characterization of individuals with 30-40 CAG repeats in the Huntington disease (HD) gene reveals HD cases with 36 repeats and apparently normal elderly individuals with 36-39 repeats. Am J Hum Genet 1996;59(1):16–22. Myers RH, Marans KS, MacDonald ME. Genetic Instabilities and Hereditary Neurological Diseases. In: Wells RD, Warren ST, eds. San Diego: Academic Press; 1998:301–23. Duyao M, Ambrose C, Myers R, et al. Trinucleotide repeat length instability and age of onset in Huntington’s disease. Nat Genet 1993;4(4):387–92. Telenius H, Kremer HP, Theilmann J, et al. Molecular analysis of juvenile Huntington disease: the major influence on (CAG)n repeat length is the sex of the affected parent. Hum Mol Genet 1993;2(10):1535–40. Wexler NS, Lorimer J, Porter J, et al. Venezuelan kindreds reveal that genetic and environmental factors modulate Huntington’s disease age of onset. Proc Natl Acad Sci U S A 2004;101(10):3498–503. Gutekunst CA, Norflus F, Hersch SM. The neuropathology of Huntington’s disease. In: Bates G, Harper PS, Jones L, eds. Huntington’s disease. Oxford: Oxford University Press; 2002:251–75.

Exploiting Yeast Genetics to Inform Therapeutic Strategies 17. Scherzinger E, Lurz R, Turmaine M, et al. Huntingtin-encoded polyglutamine expansions form amyloid-like protein aggregates in vitro and in vivo. Cell 1997;90(3):549–58. 18. Davies SW, Turmaine M, Cozens BA, et al. Formation of neuronal intranuclear inclusions underlies the neurological dysfunction in mice transgenic for the HD mutation. Cell 1997;90(3):537–48. 19. DiFiglia M, Sapp E, Chase KO, et al. Aggregation of huntingtin in neuronal intranuclear inclusions and dystropic neurites in brain. Science 1997;277(5334):1990–3. 20. Rockabrand E, Slepko N, Pantalone A, et al. The first 17 amino acids of Huntingtin modulate its sub-cellular localization, aggregation and effects on calcium homeostasis. Hum Mol Genet 2007;16(1):61–77. 21. Atwal RS, Xia J, Pinchev D, Taylor J, Epand RM, Truant R. Huntingtin has a membrane association signal that can modulate huntingtin aggregation, nuclear entry and toxicity. Hum Mol Genet 2007;16(21):2600–15. 22. Khoshnan A, Ko J, Patterson PH. Effects of intracellular expression of anti-huntingtin antibodies of various specificities on mutant huntingtin aggregation and toxicity. Proc Natl Acad Sci U S A 2002;99(2):1002–7. 23. Qin ZH, Wang Y, Sapp E, et al. Huntingtin bodies sequester vesicle-associated proteins by a polyproline-dependent interaction. J Neurosci 2004;24(1):269–81. 24. Bhattacharyya A, Thakur AK, Chellgren VM, et al. Oligoproline effects on polyglutamine conformation and aggregation. J Mol Biol 2006;355(3):524–35. 25. Dehay B, Bertolotti A. Critical role of the proline-rich region in Huntingtin for aggregation and cytotoxicity in yeast. J Biol Chem 2006;281(47):35608–15. 26. Suopanki J, Gotz C, Lutsch G, et al. Interaction of huntingtin fragments with brain membranes–clues to early dysfunction in Huntington’s disease. J Neurochem 2006;96(3):870–84. 27. Duennwald ML, Jagadish S, Muchowski PJ, Lindquist S. Flanking sequences profoundly alter polyglutamine toxicity in yeast. Proc Natl Acad Sci U S A 2006;103(29):11045–50. 28. Duennwald ML, Jagadish S, Giorgini F, Muchowski PJ, Lindquist S. A network of protein interactions determines polyglutamine toxicity. Proc Natl Acad Sci U S A 2006;103(29):11051–6. 29. Andrade MA, Bork P. HEAT repeats in the Huntington’s disease protein. Nat Genet 1995;11(2):115–6.

173

30. Harjes P, Wanker EE. The hunt for huntingtin function: interaction partners tell many different stories. Trends Biochem Sci 2003;28(8):425–33. 31. Bates GP. History of genetic disease: the molecular genetics of Huntington disease – a history. Nat Rev Genet 2005;6(10):766–73. 32. Li SH, Li XJ. Huntingtin-protein interactions and the pathogenesis of Huntington’s disease. Trends Genet 2004;20(3):146–54. 33. Outeiro TF, Muchowski PJ. Molecular genetics approaches in yeast to study amyloid diseases. J Mol Neurosci 2004;23(1– 2):49–60. 34. Outeiro TF, Giorgini F. Yeast as a drug discovery platform in Huntington’s and Parkinson’s diseases. Biotechnol J 2006;1(3):258–69. 35. Uptain SM, Lindquist S. Prions as proteinbased genetic elements. Annu Rev Microbiol 2002;56:703–41. 36. Chien P, Weissman JS, DePace AH. Emerging principles of conformation-based prion inheritance. Annu Rev Biochem 2004;73:617–56. 37. Willingham S, Outeiro TF, DeVit MJ, Lindquist SL, Muchowski PJ. Yeast genes that enhance the toxicity of a mutant huntingtin fragment or alpha-synuclein. Science 2003;302(5651):1769–72. 38. Zhang X, Smith DL, Meriin AB, et al. A potent small molecule inhibits polyglutamine aggregation in Huntington’s disease neurons and suppresses neurodegeneration in vivo. Proc Natl Acad Sci U S A 2005;102(3):892–7. 39. Giorgini F, Guidetti P, Nguyen Q, Bennett SC, Muchowski PJ. A genomic screen in yeast implicates kynurenine 3-monooxygenase as a therapeutic target for Huntington disease. Nat Genet 2005;37(5):526–31. 40. Puccio H, Koenig M. Recent advances in the molecular pathogenesis of Friedreich ataxia. Hum Mol Genet 2000;9(6):887–92. 41. Madeo F, Herker E, Wissing S, Jungwirth H, Eisenberg T, Frohlich KU. Apoptosis in yeast. Curr Opin Microbiol 2004;7(6):655–60. 42. Sokolov S, Pozniakovsky A, Bocharova N, Knorre D, Severin F. Expression of an expanded polyglutamine domain in yeast causes death with apoptotic markers. Biochim Biophys Acta 2006;1757(5–6):660–6. 43. Flower TR, Chesnokova LS, Froelich CA, Dixon C, Witt SN. Heat shock prevents alpha-synuclein-induced apoptosis in a yeast model of Parkinson’s disease. J Mol Biol 2005;351(5):1081–100. 44. Krobitsch S, Lindquist S. Aggregation of huntingtin in yeast varies with the length of the polyglutamine expansion and the expression

174

45.

46.

47.

48.

49.

50.

51.

52.

Giorgini and Muchowski of chaperone proteins. Proc Natl Acad Sci U S A 2000;97(4):1589–94. Muchowski PJ, Schaffar G, Sittler A, Wanker EE, Hayer-Hartl MK, Hartl FU. Hsp70 and hsp40 chaperones can inhibit self-assembly of polyglutamine proteins into amyloid-like fibrils. Proc Natl Acad Sci U S A 2000;97(14):7841–6. Meriin AB, Zhang X, He X, Newnam GP, Chernoff YO, Sherman MY. Huntington toxicity in yeast model depends on polyglutamine aggregation mediated by a prion-like protein Rnq1. J Cell Biol 2002;157(6):997–1004. Hughes RE, Lo RS, Davis C, et al. Altered transcription in yeast expressing expanded polyglutamine. Proc Natl Acad Sci U S A 2001;98(23):13201–6. Tam S, Geller R, Spiess C, Frydman J. The chaperonin TRiC controls polyglutamine aggregation and toxicity through subunit-specific interactions. Nat Cell Biol 2006;8(10):1155–62. Meriin AB, Zhang X, Miliaras NB, et al. Aggregation of expanded polyglutamine domain in yeast leads to defects in endocytosis. Mol Cell Biol 2003;23(21):7554–65. Meriin AB, Zhang X, Alexandrov IM, et al. Endocytosis machinery is involved in aggregation of proteins with expanded polyglutamine domains. Faseb J 2007;21(8):1915–25. Muchowski PJ, Ning K, D’Souza-Schorey C, Fields S. Requirement of an intact microtubule cytoskeleton for aggregation and inclusion body formation by a mutant huntingtin fragment. Proc Natl Acad Sci U S A 2002;99(2):727–32. Solans A, Zambrano A, Rodriguez M, Barrientos A. Cytotoxicity of a mutant huntingtin fragment in yeast involves early alterations in mitochondrial OXPHOS complexes II and III. Hum Mol Genet 2006;15(20): 3063–81.

53. Schwarcz R. The kynurenine pathway of tryptophan degradation as a drug target. Curr Opin Pharmacol 2004;4(1):12–7. 54. Panozzo C, Nawara M, Suski C, et al. Aerobic and anaerobic NAD+ metabolism in Saccharomyces cerevisiae. FEBS Lett 2002;517(1– 3):97–102. 55. Ehrnhoefer DE, Duennwald M, Markovic P, et al. Green tea (-)-epigallocatechin-gallate modulates early events in huntingtin misfolding and reduces toxicity in Huntington’s disease models. Hum Mol Genet 2006;15(18):2743–51. 56. Giaever G, Chu AM, Ni L, et al. Functional profiling of the Saccharomyces cerevisiae genome. Nature 2002;418(6896):387–91. 57. Winzeler EA, Shoemaker DD, Astromoff A, et al. Functional characterization of the S. cerevisiae genome by gene deletion and parallel analysis. Science 1999;285(5429):901–6. 58. Hansson O, Nylandsted J, Castilho RF, Leist M, Jaattela M, Brundin P. Overexpression of heat shock protein 70 in R6/2 Huntington’s disease mice has only modest effects on disease progression. Brain Res 2003;970(1–2):47–57. 59. Hay DG, Sathasivam K, Tobaben S, et al. Progressive decrease in chaperone protein levels in a mouse model of Huntington’s disease and induction of stress proteins as a therapeutic approach. Hum Mol Genet 2004;13(13):1389–405. 60. Ravikumar B, Vacher C, Berger Z, et al. Inhibition of mTOR induces autophagy and reduces toxicity of polyglutamine expansions in fly and mouse models of Huntington disease. Nat Genet 2004;36(6):585–95. 61. Botstein D, Chervitz SA, Cherry JM. Yeast as a model organism. Science 1997; 277(5330):1259–60. 62. Chopra V, Fox JH, Lieberman G, et al. A smallmolecule therapeutic lead for Huntington’s disease: preclinical pharmacology and efficacy of C2-8 in the R6/2 transgenic mouse. Proc Natl Acad Sci U S A 2007;104(42):16685–89.

Chapter 10 Global Proteomic Analysis of Saccharomyces cerevisiae Identifies Molecular Pathways of Histone Modifications Jessica Jackson and Ali Shilatifard Summary The very long DNA of the eukaryotic cells must remain functional when packaged into the cell nucleus. Although we know very little about this process, it is clear at this time that chromatin and its post-translational modifications play a pivotal role. Yeast Saccharomyces cerevisiae provides a powerful genetic and biochemical model system for deciphering the molecular machinery involved in chromatin modification and transcriptional regulation. In this chapter, we describe a novel method, the Global Proteomic analysis in S. cerevisiae (GPS), for the global analysis of the molecular machinery required for proper histone modifications. Since many of the molecular machineries involved in chromatin biology are highly conserved from yeast to humans, GPS has proven to be an outstanding method for the identification of the molecular pathways involved in chromatin modifications. Key words: Chromatin, Histones, Histone modifications, Transcription, RNA polymerase II, COMPASS, Histone methylations, Histone acetylations, Methylase, GPS

1. Introduction Chromatin is an array of nucleosomes containing 146 bp of DNA wrapped twice around an octamer composed of two copies of each of the core histone proteins H2A, H2B, H3, and H4 (1–3). Linker histones (e.g., H1) and other non-histone proteins can further pack nucleosomes to form higher order chromatin structures. Core histones are highly evolutionarily conserved proteins, with a flexible amino-terminal tail and a globular domain (2). Structural studies have demonstrated that interactions between the globular domains of each of the core histones form the nucleosome scaffold (3). These studies also confirmed that histone amino-terminal tails protrude outward from the nucleosome I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_10, © Humana Press, a part of Springer Science + Business Media, LLC 2009

175

176

Jackson and Shilatifard

and can influence nucleosome–nucleosome interactions as well as interactions between nucleosomes and regulatory factors (3). Histones can be modified by a variety of post-translational modifications such as acetylation, phosphorylation, ubiquitination, methylation, sumoylation, and ADP ribosylation (4–7). However, several novel post-translational modifications of histones have been identified recently that occur within the core region of histones (8–11). Modifications within the histone core can alter DNA–histone interactions within and between nucleosomes and thus affect higher order chromatin structures. Nucleosomes have now emerged as an active participant in the regulation of many pathways associated through chromosomal DNA. In this regard, the covalent modifications of histones can provide a combinatorial effect which can then be translated by nuclear proteins to influence a multitude of cellular processes such as transcription, replication, DNA repair, and cell cycle progression (12). Two protein families, the Trithorax group (Trx), which is the Drosophila homolog of human mixed-lineage leukemia (MLL), and the Polycomb group (Pc), are essential in the regulation of gene expression throughout development (13–15). Chromosomal translocations in the MLL gene result in hematological malignancies, including acute myeloid and lymphoid leukemia (13–16). Although these cytogenic abnormalities were discovered over 25 years ago (14), little is known about the biochemical functions of MLL, its protein complexes, its translocation partners, and, particularly, why translocations result in leukemia. The best understood target genes for MLL are the clustered homeotic Hox genes. Hox gene products are involved in the specification of cell fate. Genetic studies have demonstrated that the Trx and Pc groups of proteins cooperate to antagonistically regulate Hox gene expression. The Trx group positively regulates gene expression, while the Pc group of proteins is essential for the maintenance of the silenced state of homeotic genes. Many members of the Trx and Pc groups of proteins contain a 130–140-amino acid motif, termed the SET domain (17, 18). The SET domain is found in a variety of chromatin-associated proteins. We identified the Saccharomyces cerevisiae Set1 protein as the homolog of MLL and isolated the yeast Set1-containing protein complex. This complex is named COMPASS (Complex of Proteins Associated with Set1) (19). COMPASS was the first enzyme complex identified that catalyzes the mono-, di-, and trimethylation of H3K4 (19–22). Following the identification of yeast COMPASS, the MLL protein was also isolated in a COMPASS-like complex capable of catalyzing the methylation of H3K4 (23–26). Recent biochemical studies demonstrated that the Pc group of proteins is also histone-methyltransferase-capable of methylating lysine 27 of histone H3 (H3K27) (27–29). Together, these biochemical studies suggest the existence of a fundamental role for the regulation of the patterns of histone methylations in development.

Global Proteomic Analysis of Saccharomyces cerevisiae

177

Large-scale alterations in the histone modification landscape in chromatin accompany gene expression changes associated with differentiation and cancer. Histone modifications play a major role in the reprogramming of the genome. Histone H3K4 methylation is a hallmark of transcriptionally active genes (5). It antagonistically functions with H3K27 methylation patterns to regulate stem cell self-renewal and differentiation. Understanding how these modifications contribute to the differential regulation of the over 30,000 human genes for proper differentiation and development remains a major challenge. Much emphasis has focused on identifying the modifications and their locations within the genome; however, very little is known about the machinery that catalyzes the addition or the removal of the modifications throughout development. To better define the molecular machinery and regulatory pathways involved in the post-translational modifications of histone H3K4 methylation in yeast, we devised a proteomic screen we call Global Proteomic analysis in S. cerevisiae (GPS)(30) (Fig. 1). In GPS, we prepare extracts from each of the nonessential yeast gene deletion mutants. These extracts will then be analyzed for defects in modifications of the histones by sodium

Fig. 1. Global proteomic analysis of S. cerevisiae (GPS). Schematic representation of GPS employing a polyclonal antibody specific to lysine 4 methylated histone H3. As described in detail in the text, cell extracts from the entire yeast deletion collection were made, and these extracts were transferred to 96-well plates. Each extract was applied to a 16% SDS/ PAGE and subjected to Western analysis using antibodies towards methylated histone H3. In this figure, the Western data from a plate containing a known component of COMPASS are shown. Position E3 represents extracts from a strain deleted for the CPS50 subunit of COMPASS. Arrows at position A1 and H3 indicate empty wells as plate markers.

178

Jackson and Shilatifard

dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) and Western blotting. Antibodies specific to H3K4-methylated histones will be used as probes. Employing GPS, we have been able to identify factors that are required for proper histone H3K4 methylation by COMPASS.

2. Materials 2.1. Growth of Nonessential Gene Deletion Mutants

1. The YPD medium used for growing the yeast gene deletion mutants is composed of 1% yeast extract, 2% proteo-peptone, and 2% dextrose (Sigma catalog # G8270). The dextrose was autoclaved separately from the remaining ingredients and was added immediately after it was removed from the autoclave while the solutions were still hot. Nunc™ brand OmniTray, single-well containers were used when making agar plates. For each plate, 2% agar was added to the YP media before autoclaving, and then combined with the dextrose. 2. The cells from which the protein extracts were made were inoculated using disposable, 96-well pinning devices (Genetix catalog # X5054) onto Nunc™ brand OmniTrays (Fisher Scientific catalog # 12-565-450) containing solid YPD media from frozen stocks in which the yeast gene deletion mutants were kept in 96-well plates (Fisher Scientific catalog # 07-20090). The plates were allowed to grow for 48 h at 30 °C. From there, the cells were then inoculated into 5 ml of liquid YPD contained in sterile, polypropylene, 15-ml conical tubes, (MidSci catalog # 3018Y) and grown for 24 h at 30 °C.

2.2. Preparation of the Extracts

1. Cells were collected by centrifugation, transferred to 1.5-ml microcentrifuge tubes (SciMart catalog # GS-8509-N), and resuspended in 200 μl of lysis buffer (0.25 M Sucrose, 60 mM potassium chloride, 14 mM sodium chloride, 5 mM magnesium chloride, 1 mM calcium chloride, and 0.8% Triton X-100), and then 200 μl of 0.5 mm glass beads (BioSpec catalog # 11079105) was added. 2. After lysis and centrifugation, the supernatants were removed and the cell pellets were transferred, with water, to flexible 96-well PCR plates (Fisher Scientific catalog # 05-500-48) that were seated in Costar nonsterile 96-well flat-bottom culture plates (Fisher Scientific catalog # 07-200-90). 3. 4× Laemmli loading buffer (1 ml water, 2 ml, 1M Tris-HCl, pH 6.8, 3.2 ml 100% glycerol, 0.64 g SDS (Fisher Scientific catalog # BP166-500), 1.6 ml 2-mercaptoethanol (Sigma catalog # M7154), and 1 mg bromophenol blue) were added

Global Proteomic Analysis of Saccharomyces cerevisiae

179

to the pellets of the lysed cells, in addition to water, and then heated to 95 °C in an MJ thermocycler for 5 min. 2.3. Analysis of Extracts

1. 16% SDS gels were used to separate the extracts by PAGE, using a 1:10 dilution of the electrophoresis buffer (0.25 M Tris, 2 M glycine, 10% SDS). 2. The resolved proteins were blotted to GE™ brand 0.45-μm nitrocellulose membranes (ISC Bioexpress catalog # F-319030X3) that were submerged in Western transfer buffer (30 mM Tris, 0.15 M glycine, 20% methanol). 3. Nitrocellulose-bound proteins were probed with mono- and polyclonal antibodies specific for numerous histone modifications. Antibodies were diluted into a 1:10 dilution of Trisbuffered saline (TBS) pH 7.4 (0.1 M Tris, 1.5 M sodium chloride, 7 mM calcium chloride, 4 mM magnesium chloride). Washing of the membranes was performed in a 1:10 dilution of TBST pH 7.4 (0.1 M Tris, 1.5 M sodium chloride, 7 mM calcium chloride, 4 mM magnesium chloride, 0.5% Tween). 4. Blots were visualized by the exposure of the membranes to 14 × 17 in. IsoMax film (SciMart catalog # GX-331417) through the use of enhanced chemiluminescence (Perkin Elmer catalog # NEL104001EA).

3. Methods 1. Through the aid of a 96-well pinning device, frozen stocks from each of the approximately 4,800 nonessential yeast gene deletion mutants were inoculated onto an agar plate containing YPD, and allowed to grow for 48 h at 30 °C. The cells were then inoculated into 96 conical tubes (15 ml) containing 5 ml YPD, and grown for 24 h at 30 °C with continuous rotary movement. 2. The cells were then centrifuged at 500 g for 5 min. Using 1 ml of water, each cell pellet was washed and transferred to 1.5-ml microcentrifuge tubes. The tubes were centrifuged at 2,000 rpm for 5 min, the water was removed, and the cell pellets were resuspended in 200 μl of lysis buffer (0.25 M sucrose, 60 mM potassium chloride, 14 mM sodium chloride, 5 mM magnesium chloride, 1 mM calcium chloride, 0.8% Triton X-100), along with 200 μl of 0.5 mm glass beads. The tubes were then vigorously vortexed for 25 min at 4 °C in order to lyse the cells. 3. Using a 16G needle, cell lysates were obtained by puncturing a small hole at the bottom of the 1.5-ml tube. The tube was

180

Jackson and Shilatifard

then placed in a 15-ml conical tube to collect the contents. The 1.5-ml/15-ml tube construct was centrifuged for 10 min at 4 °C, which formed a white pellet and cloudy supernatant. The supernatant was removed and the pellet was resuspended in 100 μl of water and transferred to a 96-well PCR plate, to which 50 μl of 4× Laemmli loading buffer (1 ml water, 2 ml 1 M Tris, pH 6.8, 3.2 ml 100% glycerol, 0.64 g SDS, 1.6 ml 2-mercaptoethanol, 1 mg bromophenol blue) was added. Aluminum sealing film was placed on top of the 96-well plate, and the extracts were briefly vortexed before being heated at 95 °C for 5 min. 4. Approximately 10 μl of each extract was loaded onto a 16% SDS gel and subjected to PAGE, where the proteins were resolved at 50 mA for 1 h in an electrophoresis buffer. The resolved proteins contained in the gel were blotted onto nitrocellulose membranes submerged in a Western transfer buffer, and ran at 400 mA for 1 h. Protein-containing membranes were blocked in a milk solution consisting of non-fat dry milk and water for 30 min. 5. Antibodies specific for a variety of histone modifications were diluted in TBS and incubated with the membranes overnight at 4 °C with constant shaking. Following the removal of the antibody, the blots were subjected to three 10-min washes with TBST, after which a secondary antibody was applied and allowed to incubate with the membranes for 30 min. Three additional washes were performed in TBST. The membranes were exposed to 14 × 17 in. IsoMax film with exposure times ranging from 2 to 30 min. Through a chemiluminescent reaction, specific alterations to the histones were visualized on the film.

4. Notes For COMPASS to trimethylate histone H3 in vivo, histone H2B must be monoubiquitinated by the Rad6/Bre1 complex (5, 30–32). To identify other factors and the molecular machinery required for proper H2B monoubiquitination and, therefore, H3K4 methylation, we screened the entire yeast collection via GPS (30). Our screen revealed that several other gene products, including the subunit of the Paf1 complex (Fig. 2) (33, 34) and the subunits of the Bur1/Bur2 kinase (Fig. 3) (35), are required for proper H2B monoubiquitination and, therefore, H3K4 methylation. Furthermore, using polyclonal antibodies specifically recognizing mono-, di-, and trimethylated H3K4, we were able to demonstrate that the Cps40 subunit of COMPASS is required for proper H3K4 trimethylation by the enzyme (Fig. 4) (36).

Global Proteomic Analysis of Saccharomyces cerevisiae

181

Fig. 2. Surveying the S. cerevisiae genome with GPS identified the RNA polymerase II elongation complex, Paf1, required for the methylation of lysine 4 of histone H3. (a, b) Employing GPS, extracts of S. cerevisiae mutants missing one of the approximately 4,800 nonessential genes were tested for the presence of Lys4-methylated histone H3. Strains lacking either (a) Rtf1 or (b) Paf1 were defective for this histone modification. Short arrows at position b10 in (a), d1 in (b), and h3 in both, indicate empty wells as plate markers.

Fig. 3. Deletion of BUR2 results in a reduction of the levels of histone H3 lysine 4 and lysine 79 methylation. (a) Extracts prepared from plates containing single deletion mutants from the nonessential deletion consortium were subjected to SDS-PAGE and Western analysis using antibodies directed against dimethylated lysine 4 and lysine 79 of histone H3. Arrows indicate the lane containing the BUR2 deletion mutant. Several other sites that are null for histone methylation represent slow-growing strains or plate markers. (b) Deletion of BUR2 results in a decrease in the di- and trimethylation of H3K4 (lanes 3 and 4). After a CEN-URA3 plasmid containing HA-tagged Bur2 was introduced into the BUR2 deletion mutant, histone H3K4 methylation was restored to levels comparable to wild-type (lanes 9 and 10 ).

182

Jackson and Shilatifard

Fig. 4. Multiple subunits of COMPASS are required for histone H3 lysine 4 trimethylation. (a, b) Extracts were made from each of the nonessential gene deletion mutants of the S. cerevisiae genome. In GPS, the extracts from the collection were resolved by SDS-PAGE and tested for the presence of (a) dimethylated or (b) trimethylated lysine 4 of histone H3. Short arrows at positions A2 and H3 indicate empty wells as place markers. Long arrows indicate the position of the CPS60 deletion mutant at plate location B10. (c, d) GPS results tested for the presence of (c) dimethylated or (d) trimethylated lysine 4 of histone H3. Long arrows indicate the position of the CPS40 deletion mutant at plate location B9.

The human homolog of Cps40, the Ash2 protein, also functions similarly within mammalian H3K4 methylase complexes, indicating that the H3K4 methylation machinery is highly conserved from yeast to humans (37). Via GPS, we also determined that the Ctk complex, which is required for RNA Polymerase II CTD phosphorylation and, therefore, the rate of transcription elongation, can also regulate the pattern of H3K4 monomethylation (Fig. 5) (38). GPS has been instrumental in defining the molecular pathway required for proper H3K4 methylation by COMPASS (Fig. 6). In addition to GPS being powerful in defining the molecular pathway for H3K4 methylation, the employment of an H3K56 acetyl specific antibody was also able to identify the enzymatic machinery responsible for this histone acetylation (Fig. 7) (10). Work by other laboratories has also resulted in this same observation (39–42).

Fig. 5. Loss of CTD kinase-1 complex members affects histone H3K4 monomethylation. (a–d) GPS screen demonstrating that the deletion of either CTK1 or CTK3 results in reduced levels of H3K4 monomethylation and (a and c) but not H3K4 trimethylation (b and d).

Fig. 6. Schematic representation of the GPS identified molecular pathway required for proper histone H3 methylation on lysines 4 and 79. Employing GPS we identified in yeast initially that the Rad6/Bre1 complex is required for proper monoubiquitination of histone H2B on lysine 123. Histone H2B monoubiquitination signals for the activation of histone H3 methylation on lysines 4 and 79 by COMPASS and Dot1p, respectively. This mechanism is known as histone crosstalk. GPS has also uncovered the role of other factors such as the Paf1 complex and the Bur1/Bur2 complex in proper histone H2B monoubiquitination. The Paf1 complex can also serve as a platform for the association of COMPASS and other histone methyltransferases with the elongating form of RNA Polymerase II (Pol II). The monoubiquitinated histone H2B is deubiquitinated by the action of deubiquitinating enzymes such as Ubp8 and Ubp10. The enzymatic machinery involved in the histone crosstalk pathway is highly conserved from yeast to humans.

184

Jackson and Shilatifard

Fig. 7. Global proteomic analyses in defining factors required for proper H3K56 acetylation. (a–c) Total cellular extracts were prepared from each of the nonessential yeast gene deletion mutants of the S. cerevisiae genome. Each extract was subjected to SDS-PAGE, Western blotted, and probed with the anti-acetyl H3 Lys56 antibody. Long arrows indicate the positions of (a) spt10D in which H3K56 acetylation is present, (b) asf1D, and (c) rtt109D where H3K56 acetylation is absent. Short arrows indicate empty wells as plate markers. (d) Extracts from the rtt109D strain were titrated and analyzed by SDS-PAGE to determine the level of the H3K56 loss in the absence of Rtt109. Western blots probed with the appropriate antibodies demonstrated a loss in H3K56 acetylation. H3 Lys4 trimethylation remained unaffected and anti-acetyl Histone H3 was used as a loading control.

GPS has most certainly been a very successful method for identifying the molecular pathway required for proper H3K4 methylation in yeast (Fig. 6). The information obtained from this screen has also proven to be very useful as a template for identifying factors involved in H3K4 methylation in higher eukaryotic organisms. This method is a testimony to the great value of the S. cerevisiae deletion collection. One disadvantage of GPS is the requirement for an antiserum sensitive enough to detect the relatively small amounts of protein made available by our method. To surmount this problem, we are now developing methods that do not require an antibody specific for the given post-translational modifications. Having said this, GPS can also be used for defining the molecular machinery required for other post-translational modification pathways such as phosphorylation, acetylation, and glycosylation when antibodies capable of identifying such modifications on proteins in yeast S. cerevisiae are available.

Global Proteomic Analysis of Saccharomyces cerevisiae

185

References 1. Kornberg, R. D. 1974. Chromatin structure: a repeating unit of histones and DNA. Science 184:868–71. 2. Kornberg, R. D., and Y. Lorch. 1999. Twentyfive years of the nucleosome, fundamental particle of the eukaryote chromosome. Cell 98:285–94. 3. Luger, K., A. W. Mader, R. K. Richmond, D. F. Sargent, and T. J. Richmond. 1997. Crystal structure of the nucleosome core particle at 2.8 A resolution. Nature 389:251–60. 4. Berger, S. L. 2007. The complex language of chromatin regulation during transcription. Nature 447:407–12. 5. Shilatifard, A. 2006. Chromatin modifications by methylation and ubiquitination: implications in the regulation of gene expression. Annu Rev Biochem 75:243–69. 6. Workman, J. L., and R. E. Kingston. 1998. Alteration of nucleosome structure as a mechanism of transcriptional regulation. Annu Rev Biochem 67:545–79. 7. Wu, J., and M. Grunstein. 2000. 25 years after the nucleosome model: chromatin modifications. Trends Biochem Sci 25:619–23. 8. Lacoste, N., R. T. Utley, J. M. Hunter, G. G. Poirier, and J. Cote. 2002. Disruptor of telomeric silencing-1 is a chromatin-specific histone H3 methyltransferase. J Biol Chem 277:30421–4. 9. Ng, H. H., Q. Feng, H. Wang, H. ErdjumentBromage, P. Tempst, Y. Zhang, and K. Struhl. 2002. Lysine methylation within the globular domain of histone H3 by Dot1 is important for telomeric silencing and Sir protein association. Genes Dev 16:1518–27. 10. Schneider, J., P. Bajwa, F. C. Johnson, S. R. Bhaumik, and A. Shilatifard. 2006. Rtt109 is required for proper H3K56 acetylation: a chromatin mark associated with the elongating RNA polymerase II. J Biol Chem 281:37270–4. 11. van Leeuwen, F., P. R. Gafken, and D. E. Gottschling. 2002. Dot1p modulates silencing in yeast by methylation of the nucleosome core. Cell 109:745–56. 12. Bhaumik, S. R., E. Smith, and A. Shilatifard. 2007. Covalent modifications of histones during development and disease pathogenesis. Nat Struct Mol Biol 14:1008–16. 13. Hess, J. L. 2004. MLL: a histone methyltransferase disrupted in leukemia. Trends Mol Med 10:500–7. 14. Rowley, J. D. 1998. The critical role of chromosome translocations in human leukemias. Annu Rev Genet 32:495–519.

15. Tenney, K., and A. Shilatifard. 2005. A COMPASS in the voyage of defining the role of trithorax/MLL-containing complexes: linking leukemogensis to covalent modifications of chromatin. J Cell Biochem 95:429–36. 16. Tenney, K., M. Gerber, A. Ilvarsonn, J. Schneider, M. Gause, D. Dorsett, J. C. Eissenberg, and A. Shilatifard. 2006. Drosophila Rtf1 functions in histone methylation, gene expression, and Notch signaling. Proc Natl Acad Sci U S A 103:11970–4. 17. Jones, R. S., and W. M. Gelbart. 1993. The Drosophila Polycomb-group gene enhancer of zeste contains a region with sequence similarity to trithorax. Mol Cell Biol 13:6357–66. 18. Tschiersch, B., A. Hofmann, V. Krauss, R. Dorn, G. Korge, and G. Reuter. 1994. The protein encoded by the Drosophila positioneffect variegation suppressor gene Su(var)3–9 combines domains of antagonistic regulators of homeotic gene complexes. EMBO J 13:3822–31. 19. Miller, T., N. J. Krogan, J. Dover, H. Erdjument-Bromage, P. Tempst, M. Johnston, J. F. Greenblatt, and A. Shilatifard. 2001. COMPASS: a complex of proteins associated with a trithorax-related SET domain protein. Proc Natl Acad Sci U S A 98:12902–7. 20. Krogan, N. J., J. Dover, S. Khorrami, J. F. Greenblatt, J. Schneider, M. Johnston, and A. Shilatifard. 2002. COMPASS, a histone H3 (lysine 4) methyltransferase required for telomeric silencing of gene expression. J Biol Chem 277:10753–5. 21. Nagy, P. L., J. Griesenbeck, R. D. Kornberg, and M. L. Cleary. 2002. A trithorax-group complex purified from Saccharomyces cerevisiae is required for methylation of histone H3. Proc Natl Acad Sci U S A 99:90–4. 22. Roguev, A., D. Schaft, A. Shevchenko, W. W. Pijnappel, M. Wilm, R. Aasland, and A. F. Stewart. 2001. The Saccharomyces cerevisiae Set1 complex includes an Ash2 homologue and methylates histone 3 lysine 4. EMBO J 20:7137–48. 23. Cho, Y. W., T. Hong, S. Hong, H. Guo, H. Yu, D. Kim, T. Guszczynski, G. R. Dressler, T. D. Copeland, M. Kalkum, and K. Ge. 2007. PTIP associates with MLL3- and MLL4-containing histone H3 lysine 4 methyltransferase complex. J Biol Chem 282(28):20395–406. 24. Hughes, C. M., O. Rozenblatt-Rosen, T. A. Milne, T. D. Copeland, S. S. Levine, J. C. Lee, D. N. Hayes, K. S. Shanmugam, A. Bhattacharjee, C. A. Biondi, G. F. Kay, N. K. Hayward, J. L. Hess, and M. Meyerson. 2004.

186

25.

26.

27.

28.

29.

30.

31.

32.

33.

Jackson and Shilatifard Menin associates with a trithorax family histone methyltransferase complex and with the hoxc8 locus. Mol Cell 13:587–97. Lee, J. H., and D. G. Skalnik. 2005. CpGbinding protein (CXXC finger protein 1) is a component of the mammalian Set1 histone H3-Lys4 methyltransferase complex, the analogue of the yeast Set1/COMPASS complex. J Biol Chem 280:41725–31. Lee, J. H., C. M. Tate, J. S. You, and D. G. Skalnik. 2007. Identification and characterization of the human Set1B histone H3-Lys4 methyltransferase complex. J Biol Chem 282:13419–28. Barski, A., S. Cuddapah, K. Cui, T. Y. Roh, D. E. Schones, Z. Wang, G. Wei, I. Chepelev, and K. Zhao. 2007. High-resolution profiling of histone methylations in the human genome. Cell 129:823–37. Cao, R., L. Wang, H. Wang, L. Xia, H. Erdjument-Bromage, P. Tempst, R. S. Jones, and Y. Zhang. 2002. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science 298:1039–43. Lee, T. I., R. G. Jenner, L. A. Boyer, M. G. Guenther, S. S. Levine, R. M. Kumar, B. Chevalier, S. E. Johnstone, M. F. Cole, K. Isono, H. Koseki, T. Fuchikami, K. Abe, H. L. Murray, J. P. Zucker, B. Yuan, G. W. Bell, E. Herbolsheimer, N. M. Hannett, K. Sun, D. T. Odom, A. P. Otte, T. L. Volkert, D. P. Bartel, D. A. Melton, D. K. Gifford, R. Jaenisch, and R. A. Young. 2006. Control of developmental regulators by Polycomb in human embryonic stem cells. Cell 125:301–13. Dover, J., J. Schneider, M. A. Tawiah-Boateng, A. Wood, K. Dean, M. Johnston, and A. Shilatifard. 2002. Methylation of histone H3 by COMPASS requires ubiquitination of histone H2B by Rad6. J Biol Chem 277:28368–71. Sun, Z. W., and C. D. Allis. 2002. Ubiquitination of histone H2B regulates H3 methylation and gene silencing in yeast. Nature 418:104–8. Wood, A., N. J. Krogan, J. Dover, J. Schneider, J. Heidt, M. A. Boateng, K. Dean, A. Golshani, Y. Zhang, J. F. Greenblatt, M. Johnston, and A. Shilatifard. 2003. Bre1, an E3 ubiquitin ligase required for recruitment and substrate selection of Rad6 at a promoter. Mol Cell 11:267–74. Krogan, N. J., J. Dover, A. Wood, J. Schneider, J. Heidt, M. A. Boateng, K. Dean, O. W. Ryan, A. Golshani, M. Johnston, J. F. Greenblatt, and A. Shilatifard. 2003. The Paf1 complex is required for histone H3 methylation by COMPASS and Dot1p: linking transcriptional

34.

35.

36.

37.

38.

39.

40.

41.

42.

elongation to histone methylation. Mol Cell 11:721–9. Wood, A., J. Schneider, J. Dover, M. Johnston, and A. Shilatifard. 2003. The Paf1 complex is essential for histone monoubiquitination by the Rad6-Bre1 complex, which signals for histone methylation by COMPASS and Dot1p. J Biol Chem 278:34739–42. Wood, A., J. Schneider, J. Dover, M. Johnston, and A. Shilatifard. 2005. The Bur1/Bur2 complex is required for histone H2B monoubiquitination by Rad6/Bre1 and histone methylation by COMPASS. Mol Cell 20:589–99. Schneider, J., A. Wood, J. S. Lee, R. Schuster, J. Dueker, C. Maguire, S. K. Swanson, L. Florens, M. P. Washburn, and A. Shilatifard. 2005. Molecular regulation of histone H3 trimethylation by COMPASS and the regulation of gene expression. Mol Cell 19:849–56. Steward, M. M., J. S. Lee, A. O’Donovan, M. Wyatt, B. E. Bernstein, and A. Shilatifard. 2006. Molecular regulation of H3K4 trimethylation by ASH2L, a shared subunit of MLL complexes. Nat Struct Mol Biol 13:852–4. Wood, A., A. Shukla, J. Schneider, J. S. Lee, J. D. Stanton, T. Dzuiba, S. K. Swanson, L. Florens, M. P. Washburn, J. Wyrick, S. R. Bhaumik, and A. Shilatifard. 2007. Ctk complexmediated regulation of histone methylation by COMPASS. Mol Cell Biol 27:709–20. Collins, S. R., K. M. Miller, N. L. Maas, A. Roguev, J. Fillingham, C. S. Chu, M. Schuldiner, M. Gebbia, J. Recht, M. Shales, H. Ding, H. Xu, J. Han, K. Ingvarsdottir, B. Cheng, B. Andrews, C. Boone, S. L. Berger, P. Hieter, Z. Zhang, G. W. Brown, C. J. Ingles, A. Emili, C. D. Allis, D. P. Toczyski, J. S. Weissman, J. F. Greenblatt, and N. J. Krogan. 2007. Functional dissection of protein complexes involved in yeast chromosome biology using a genetic interaction map. Nature 446:806–10. Driscoll, R., A. Hudson, and S. P. Jackson. 2007. Yeast Rtt109 promotes genome stability by acetylating histone H3 on lysine 56. Science 315:649–52. Han, J., H. Zhou, B. Horazdovsky, K. Zhang, R. M. Xu, and Z. Zhang. 2007. Rtt109 acetylates histone H3 lysine 56 and functions in DNA replication. Science 315:653–5. Tsubota, T., C. E. Berndsen, J. A. Erkmann, C. L. Smith, L. Yang, M. A. Freitas, J. M. Denu, and P. D. Kaufman. 2007. Histone H3-K56 acetylation is catalyzed by histone chaperone-dependent complexes. Mol Cell 25:703–12.

Chapter 11 Systematic Characterization of the Protein Interaction Network and Protein Complexes in Saccharomyces cerevisiae Using Tandem Affinity Purification and Mass Spectrometry Mohan Babu, Nevan J. Krogan, Donald E. Awrey, Andrew Emili, and Jack F. Greenblatt Summary Defining protein complexes is a vital aspect of cell biology because cellular processes are often carried out by stable protein complexes and their characterization often provides insights into their function. Accurate identification of the interacting proteins in macromolecular complexes is easiest after purification to near homogeneity. To this end, the tandem affinity purification (TAP) system with subsequent protein identification by high-throughput mass spectrometry was developed (1, 2) to systematically characterize native protein complexes and transient protein interactions under near-physiological conditions. The TAP tag containing two adjacent affinity purification tags (calmodulin-binding peptide and Staphylococcus aureus protein A) separated by a tobacco etch virus (TEV) protease cleavage site is fused with the open reading frame of interest. Using homologous recombination, a fusion library was constructed for the yeast Saccharomyces cerevisiae (3) in which the carboxy-terminal end of each predicted open reading frame is individually tagged in the chromosome so that the resulting fusion proteins are expressed under the control of their natural promoters (3). In this chapter, an optimized protocol for systematic protein purification and subsequent mass spectrometry-based protein identification is described in detail for the protein complexes of S. cerevisiae (4–6). Key words: Saccharomyces cerevisiae, TAP tagging, Affinity purification, Mass spectrometry, LC-MS, MALDI-TOF, Protein complex

1. Introduction Since protein interactions are important for most cellular processes, and multi-subunit protein machines actively participate in most or all such processes (5–7), it is of immense significance to I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_11, © Humana Press, a part of Springer Science + Business Media, LLC 2009

187

188

Babu et al.

systematically characterize protein–protein interactions (PPIs) in various organisms. In the budding yeast Saccharomyces cerevisiae, large-scale PPI networks were first generated by systematically using the yeast two-hybrid technique (8, 9). Subsequently, the tandem affinity purification (TAP) approach was developed for the purification of native yeast protein complexes (2, 4). In this approach, TAP-tagged proteins are expressed under normal physiological conditions, purified ~106-fold via a two-step enrichment procedure, and then characterized by mass spectrometry. Because it is relatively simple to fuse affinity tags with target proteins, this approach has also been successfully applied to many other evolutionarily diverse organisms, including Caenorhabditis elegans, Drosophila melanogaster, Escherichia coli, mammalian cells, and plants (10–15). Exploring PPI networks on a large scale provides not only additional information about well-characterized proteins but also a rational framework for elucidating the biological functions of uncharacterized proteins based on the concept of “guilt by association” (16). The TAP tag consists of two affinity purification tags, a calmodulin-binding peptide (CBP) and Staphylococcus aureus protein A, separated by a tobacco etch virus (TEV) protease cleavage site (2), to allow highly selective two-stage protein enrichment and to help reduce the number of nonspecifically bound proteins (Fig. 1a). Each tagged protein is purified first by binding to beads containing immobilized IgG and subsequently by binding to beads containing immobilized calmodulin. To facilitate protein analyses on a global scale, Ghaemmaghami et al. (3) used homologous recombination to integrate a TAP-tag in-frame immediately after the stop codon of each predicted open reading frame in its natural chromosomal location in S. cerevisiae. The TAP-tagged fusion proteins are then expressed under the control of their natural promoters, and the abundance of each TAP-tagged fusion protein was assessed by quantitative Western blotting using an antibody that binds to the protein A component of the TAP tag (3). The entire collection of TAP-tagged yeast strains has been made commercially available for academic use (http://www.openbiosystems.com/GeneExpression/Yeast/TAP/). We purified a large number of the TAP-tagged bait proteins from 2 L yeast cultures under native conditions (6). To increase the interactome coverage and confidence, the identities of the co-purifying proteins (preys) were determined using two complementary, highly sensitive mass spectrometry methods: gel-free liquid chromatography-tandem mass spectrometry (LC-MS/MS), and gel-based peptide mass fingerprinting using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). We attempted to purify 4,562 different soluble proteins, and an extensive PPI network containing high-confidence protein interactions generated from this

a C-terminal TAP-tagged fusion protein

C

C C C

Bait protein

CBP

TEV

Bait protein

Protein A

C

CBP N

Bait protein

CBP

TEV

Protein A

Second affinity Calmodulin CBP Sepharose beads purification step

CBP

Calmodulin Binding Peptide

TEV

TEV protease cleavage site

C

Contaminants

Associated proteins

Native elution (EGTA)

Bait protein

C

TAP tag

IgG Sepharose First affinity purification step beads

TEV protease cleavage

Bait protein

TEV Protein A

TEV protease cleavage

CBP

b Purified TAP-tagged Silver staining of Protein bands excised from silver Samples to be subjected SDS-polyacrylamide gel fusion protein stained SDS-polyacrylamide gel to SDS-PAGE and MALDI-TOF MS

Bait protein

Protein 1

CBP

Protein 2

Protein 3

Samples to be subjected to LC-MS/MS

Protein 4 Protein 5 Protein 6 Protein 7 Protein 8

Purified protein digested with trypsin

Protein identification

LC-MS/MS

MALDI-TOF MS

Digestion of the excised protein bands with trypsin in a 96-well plate

Mass spectrum searched against the yeast protein database S. Cerevisiae Protein Sequence Database

Fig. 1. Schematic representation of the procedures involved in the identification of protein complexes using the tandem affinity purification (TAP) approach followed by mass spectrometry analyses. The TAP-tag structure and an overview of the TAP strategy are shown in (a). Identification of the co-purifying proteins is performed using two complementary mass spectrometry methods (b). First, the purified TAP-tagged fusion protein and associated proteins are separated by SDS-PAGE followed by silver staining. The excised protein bands are digested with trypsin, and the resulting tryptic peptides are analyzed by peptide mass finger printing using MALDI-TOF MS. A yeast protein database is then searched using the Knexus software. Second, the purified TAP-tagged fusion proteins and associated proteins are digested with trypsin and analyzed by gel-free liquid chromatographytandem mass spectrometry (LC-MS/MS). Searches are performed with the tandem mass spectra of the complete S. cerevisiae protein database using the SEQUEST computer algorithm to generate probable identifications of the proteins.

190

Babu et al.

study was published in 2006 (6) at the same time as Gavin et al. (5). These authors also used the TAP method and MALDITOF MS to elucidate a similarly extensive set of yeast PPI. More recently, the protein interaction data generated from these two large-scale studies were amalgamated into a single, more reliable collection of experimentally based PPIs using a novel purification enrichment (PE) scoring system (17), and this larger, amalgamated set of PPI was used in conjunction with a Markov clustering algorithm to organize the S. cerevisiae proteome into a large number of more reliable protein complexes (18). Having attempted to purify all the yeast proteins predicted to be soluble, we are now purifying the many yeast proteins predicted to be membrane-associated (see Note 1). Purification of membrane proteins often poses unique challenges (19–22) because they are often not solubilized by the extraction buffer that we normally use for the TAP method. With some modifications of our standard procedures and with the addition of various nonionic detergents to our buffers, we are able to solubilize and purify the majority of the yeast membrane proteins. Currently, we are attempting to purify the ~1,600 S. cerevisiae proteins that are predicted to contain signal peptides or with at least one transmembrane domain. Strains for the vast majority of these purifications are already present in the collection of yeast TAP-tagged strains (3). As of March 2009, more than 1,200 yeast membrane proteins had been purified using at least two different detergents. In this chapter, detailed procedures are described for the purification of TAP-tagged bait proteins and subsequent identification of interacting protein partners by mass spectrometry. This method has been optimized for a well-characterized yeast laboratory strain and can be used to purify the subunits of highor low-abundance yeast protein complexes. This basic approach can potentially be adapted for use in other organisms.

2. Materials 2.1. Purification of TAP-Tagged Proteins from S. Cerevisiae Strains 2.1.1. Culturing TAPTagged S. Cerevisiae Strains and Cell Lysis

1. S. cerevisiae strains used for culturing in which a bait protein is TAP-tagged have been described (3). 2. Yeast extract–peptone–dextrose (YEPD) medium: Dissolve 10 g of yeast extract (Bioshop), 20 g of peptone (Bioshop), and 20 g of glucose in 800 ml of distilled water. Autoclave the media on a liquid cycle for 30 min. 3. 2 M Tris–HCl (pH 7.9) stock solution: Dissolve 242.2 g of Tris base in 800 ml of double distilled water, adjust to pH 7.9 with concentrated HCl, and make up the volume to 1 L.

Systematic Characterization of the Protein Interaction Network

191

4. 5 M NaCl stock solution: Dissolve 292.2 g of NaCl in 1 L of double-distilled water. 5. 0.5 M ethylenediaminetetraacetic acid (EDTA) (pH 8.0) stock solution: Dissolve 186.1 g of EDTA in 800 ml of double-distilled water, adjust the pH to 8.0 using 1 M NaOH, and make up the volume to 1 L. 6. HEPES-KOH (pH 7.9) stock solution: Dissolve 238.3 g of HEPES in 800 ml of double-distilled water, adjust the pH to 7.9 using 5 M KOH, and make up the volume to 1 L. 7. 1 M CaCl2: Dissolve 110.98 g of CaCl2 in 800 ml of doubledistilled water and make up the volume to 1 L. 8. 2 M KCl stock solution: Dissolve 149.1 g of KCl in 800 ml of double-distilled water and make up the volume to 1 L. 9. EDTA-free protease inhibitor tablets (Roche). Make sure to prepare YEB buffer with and without the EDTA-free protease inhibitor tablet. 10. YEB buffer with protease inhibitor: Dissolve one tablet of protease inhibitor in 50 ml YEB buffer. 11. Yeast extract buffer (YEB): Mix 36.6 ml of 2 M KCl, 3 ml of 0.5 M EDTA, 3 ml of 0.5 M ethyleneglycoltetraacetic acid (EGTA)-KOH (pH 7.9), and 30 ml of 1 M HEPES-KOH (pH 7.9) in 229 ml of double-distilled water. Store the solution at room temperature. Prior to use, add 750 ml of 1 M dithiothreitol (DTT; Bioshop) into the YEB buffer. 12. Dialysis buffer (DB): Mix 80 ml of 5 M NaCl, 20 ml of 2 M Tris–HCl (pH 7.9), 1.6 ml of 0.5 M EDTA, 800 ml of 100% glycerol, and 0.32 g of DTT with 3098.4 ml of doubledistilled water. Store the solution at 4°C. 2.1.2. Purification of TAP-Tagged Proteins

1. Store 10% Triton X-100 solution (Sigma Aldrich) at 4°C. 2. IPP buffer: Add 50 ml of 2 M Tris–HCl (pH 7.9), 200 ml of 5 M NaCl, and 100 ml of 10% Triton X-100 to 9.65 ml of sterile distilled water. 3. TEV protease cleavage buffer: Add 250 ml of 2 M Tris–HCl (pH 7.9), 4 ml of 0.5 M EDTA, 200 ml of 5 M NaCl, 100 ml of 10% Triton X-100, and 10 ml of 1 M DTT to 9.7 ml of sterile distilled water. 4. Calmodulin binding buffer: Add 200 ml of 2 M Tris–HCl (pH 7.9), 800 ml of 5 M NaCl, 80 ml of 1 M CaCl2, 400 ml of 10% Triton X-100, and 28 ml of b-mercaptoethanol solution to 38.9 ml of sterile distilled water. 5. Calmodulin wash buffer: Add 50 ml of 2 M Tris–HCl (pH 7.9), 200 ml of 5 M NaCl, 1 ml of 1 M CaCl2, 100 ml of 10% Triton X-100, and 7 ml of b-mercaptoethanol solution to 9.6 ml of sterile distilled water.

192

Babu et al.

6. Calmodulin elution buffer: Add 50 ml of 2 M Tris–HCl (pH 7.9), 200 ml of 5 M NaCl, 60 ml of 0.5 M EGTA, 7 ml of b-mercaptoethanol, and 100 ml of 10% Triton X-100 to 8.8 ml of sterile distilled water. 2.2. Protein Identification by MALDI-TOF Mass Spectrometry 2.2.1. SDS-Polyacrylamide Gel Electrophoresis (SDS-PAGE)

1. SDS stock solution (10% (w/v)): 10 g SDS is dissolved in 100 ml double-distilled water. The solution is stored at room temperature. 2. Electrode running buffer (5×): (125 mM Tris–HCl, 960 mM glycine, 0.5% (w/v) SDS, pH 8.3). 30 g of Tris–HCl, 144 g of glycine, and 10 g of SDS are dissolved in 2 L of doubledistilled water without pH adjustment and stored at 4°C. 3. Acrylamide: 30% acrylamide monomer, 0.8% N,N¢methylenebis-acrylamide (Bio-Rad). The solution is filtered through Whatman No.1 filter paper (Fischer Scientific) and stored at 4°C in the dark. 4. Separating buffer (4×): 1.5 M Tris–HCl (pH 8.7), 0.4% SDS. Store at room temperature. 5. Stacking buffer (5×): 0.5 M Tris–HCl (pH 6.8), 0.4% SDS. Store at room temperature. 6. Ammonium persulfate (10% APS): Mix 1 g of APS (Bioshop) in 10 ml of double-distilled water. Aliquot the solution into several microcentrifuge tubes in a volume of 200 ml. Store the aliquots at −20°C. 7. Water saturated n-butanol: Equal volumes of n-butanol (Sigma Aldrich) and double-distilled water are mixed in a glass bottle and left at room temperature to separate. Use the topmost layer containing n-butanol saturated with water. Store the solution at room temperature. 8. Resolving gel (12.5% polyacrylamide): Mix 7.2 ml of 4× separating buffer with 12 ml acrylamide solution, 9.6 ml double-distilled water, 120 ml 10% APS solution, and 20 ml tetramethylethylenediamine (TEMED, Bioshop). 9. Stacking gel (4.5% polyacrylamide): Mix 4 ml of 5× stacking buffer with 2.4 ml acrylamide solution, 9.6 ml double distilled water, 150 ml 10% APS solution, and 15 ml TEMED. 10. Sample stock buffer (3×): Mix 1.25 ml of 0.5 M Tris–HCl (pH 6.8), 1 ml of 100% glycerol, 1 ml of 10% SDS (w/v), and 20 mg of bromophenol blue (Bio-Rad). 11. 2× sample buffer: Add 6 ml of 3× sample stock buffer, 0.6 ml of b-mercaptomethanol, and 3.4 ml of distilled water.

2.2.2. Silver-Staining the SDS-Polyacrylamide Gel

1. Fixer: 50% methanol and 10% acetic acid (AA) are added to 400 ml sterile distilled water. 2. Sensitizer: 20 mg of fresh sodium thiosulfate (Sigma) is dissolved in 1,000 ml distilled water.

Systematic Characterization of the Protein Interaction Network

193

3. Silver nitrate solution: 2 g of silver nitrate (Fischer Scientific) is dissolved in 1,000 ml distilled water. 4. Developing solution: 1.4 ml of 37% formaldehyde and 30 g of sodium carbonate are added to 1,000 ml of distilled water. 5. Stop solution: 5 ml AA is diluted in 500 ml of distilled water. 2.2.3. Gel-Based MALDITOF Mass Spectrometry

1. An ULTRAFlex II MALDI-TOF instrument for acquiring spectral data from in-gel trypsin digested yeast samples (Bruker Daltonics, Billerica, MA). 2. MALDI target plate (Bruker). 3. Bulk C18 reverse phase resin (Sigma). 4. Knexus automation, a Windows-based program from Genomics Solutions Bioinformatics for database searches. 5. a-Cyano-4-hydroxycinnamic acid matrix solution (Fluka Buchs SG, Switzerland). 6. 1 mM HCl: Add 1 ml of 10 N HCl into 10 ml HPLC-grade water. Make a fresh solution each day and store on ice until use. 7. Trypsin stock solution: Dissolve 100 mg Boehringer Mannheim unmodified sequencing-grade trypsin (Roche) in 1 ml of 1 mM HCl. Store the trypsin stock solution at −80°C. 8. Digestion buffer: Mix 9.12 ml of 100 mM NH4HCO3, 9.12 ml of HPLC-grade water, 960 ml of 1% CaCl2, and 1 ml of trypsin stock solution. Prepared a fresh solution each day and store on ice until use. 9. 100 mM NH4HCO3: Dissolve 0.79 g NH4HCO3 in 100 ml HPLC-grade water. 10. 100 mM NH4HCO3 containing 10 mM DTT: Add 100 ml of 1 M DTT to 9.9 ml of 100 mM NH4HCO3. Prepare a fresh solution each day and store in an amber-colored bottle at room temperature in the dark until use. 11. 100 mM NH4HCO3 containing 55 mM iodoacetamide (Sigma): Add 0.103 g of iodoacetamide to 10 ml of 100 mM NH4HCO3. Prepare a fresh solution every day and store in an amber-colored bottle at room temperature in the dark until use. 12. 66% Acetonitrile (ACN), 1% AA: Mix 66 ml HPLC-grade ACN (Sigma), 33 ml of HPLC-grade water, and 1 ml AA. Store the solution in a glass bottle for up to 1 month at room temperature. 13. 75% ACN, 1% AA: Mix 75 ml HPLC-grade ACN, 24 ml HPLC-grade water, and 1 ml AA. Store the solution in a glass bottle for up to 1 month at room temperature.

194

Babu et al.

14. 2% ACN, 1% AA: Mix 2 ml of HPLC-grade ACN, 97 ml HPLC-grade water, and 1 ml AA. Store the solution in a glass bottle for up to 1 month at room temperature. 15. 1% CaCl2: Dissolve 1 g of CaCl2 in HPLC-grade water to a final volume of 100 ml. Store the solution in a glass bottle for up to 1 month at room temperature. 16. 0.1% Trifluoroacetic acid (TFA) stock solution: Dilute 0.25 ml of TFA in 250 ml of HPLC-grade water. 17. Peptide calibration standard (Bruker) solution: Lyophilized peptide standard is dissolved with 125 ml of 0.1% HPLCgrade TFA. 2.3. Protein Identification by Liquid Chromatography Mass Spectrometry (LC-MS/MS)

1. LTQ tandem mass spectrometer (Finnigan Corp, San Jose, CA, USA) to run the samples, and XCalibur software to acquire tandem mass spectra and to control the instrument. 2. Digestion buffr: Mix 599 ml of 50 mM NH4HCO3 and 1 ml of 1 M CaCl2. Store the solution at 4°C prior to use. 3. Immobilized trypsin solution: Mix 18.7 ml of digestion buffer, 1.8 ml PIERCE immobilized trypsin beads (PIERCE), 0.9 ml immobilized trypsin beads (Applied Biosciences), and 0.06 ml of 1 M CaCl2. Make sure that the pH of the immobilized trypsin solution is ~8.0. 4. 150 mm fused silica (Polymicro Technologies, Pheoenix, AZ, USA). 5. C18 reverse phase packing material (Zorbax eclipse XDBC18 resin; Agilent Technologies, Mississauga, ON, Canada). 6. Solvent A: 5% ACN, 0.5% AA, and 0.02% heptafluorobutyric acid (HFBA) 7. Solvent B: 100% ACN 8. Proxeon nano HPLC pump (Proxeon Biosciences).

3. Methods 3.1. Purification of S. Cerevisiae TAP-Tagged Fusion Proteins 3.1.1. Culturing TAP-Tagged S. Cerevisiae Strains and Cell Lysis

1. Inoculate a loop of a glycerol stock of a TAP-tagged S. cerevisiae strain (3) into 10 ml YEPD liquid medium in a 20-ml sterile culture tube. 2. Grow the culture overnight and the next day at 30°C with shaking at 250 rpm until the OD600 reaches 1.0–1.5. 3. Inoculate 10 ml of the overnight culture into 2 L fresh YEPD liquid medium in a 4-L flask (see Note 2). The culture is grown overnight at 30°C with shaking at 250 rpm until the OD600 reaches ~1.0–1.5.

Systematic Characterization of the Protein Interaction Network

195

4. Transfer 2 L S. cerevisiae culture from the shaker to clean 1-L centrifugation bottles. 5. Centrifuge the bottles containing the S. cerevisiae culture in a Beckman J-20XP Avanti centrifuge at 3,993g for 5 min at 4°C. 6. Discard the supernatant and remove excess liquid by inverting the bottles on paper towels. Keep the centrifugation bottles on ice. 7. Add 10 ml cold distilled water to the centrifugation bottles and resuspend the S. cerevisiae cell pellets using a clean 25-ml pipette. 8. Transfer the cell lysates into 50-ml polypropylene Falcon tubes. 9. Centrifuge the Falcon tubes containing the cell lysates in a Beckman J-20XP Avanti centrifuge at 3,993g for 5 min at 4°C. 10. Decant the cold water from the Falcon tubes, and resuspend the cell pellets with an equal volume of cold YEB buffer without protease inhibitor using a clean 25-ml pipette. 11. Centrifuge the Falcon tubes containing the cell lysates in an Eppendorf centrifuge 5810R using an A-4–62 rotor at 751 × g (4,000 rpm) for 5 min at 4°C. 12. Add an equal volume of cold YEB buffer with protease inhibitor to the Falcon tubes containing the cell pellets and resuspend the pellets using a clean 25-ml pipette. 13. Repeat step 11. 14. Snap-freeze the Falcon tubes containing the cell pellets using liquid nitrogen. The frozen cell pellets are stored at −80°C for future use. 15. The Falcon tubes containing the frozen cell pellets are removed from the freezer and wrapped in several layers of paper towels. 16. Using a hammer, smash the Falcon tubes containing the frozen cell pellets into small pieces. 17. Transfer each yeast cell pellet (7–10 g) into a prechilled Krups coffee grinder (Krups, Model 203–70). Avoid transferring the broken pieces of plastic. 18. Add 25% dry ice into the coffee grinder containing 7–10 g yeast cell pellets. 19. Perform lysis by grinding the yeast cell pellets to a fine powder with dry ice using the Krups coffee grinder (see Note 3). This takes approximately 2–3 min. 20. Scrape the lysed powder into a 25-ml ultracentrifuge tube (25 × 89 mm, Beckman, Part No. 355642) placed on dry ice.

196

Babu et al.

21. Resuspend the lysed powder by adding an equal volume of YEB buffer containing protease inhibitor into the 25-ml ultracentrifuge tubes (see Note 4). 22. Centrifuge the 25-ml ultracentrifuge tubes containing the cell lysates in a Beckman L8-M ultracentrifuge using a Type 70Ti rotor at 208,429 × g (45,000 rpm) for 1 h at 4°C. 23. The supernatants from the ultracentrifuge tubes are collected and transferred to dialysis tubes (Spectra/Por®, 29 mm in diameter). The supernatants are dialyzed against 4 L dialysis buffer containing 100 mM NaCl for 3 h at 4°C (see Notes 5 and 6). 24. After dialysis, the extracts are transferred from the dialysis tubes to sterile 25-ml ultracentrifuge tubes. 25. Centrifuge the 25-ml ultracentrifuge tubes containing the extracts in a Beckman L8-M ultracentrifuge using a Type 70Ti rotor at 208,429 × g (45,000 rpm) for 30 min at 4°C to remove any precipitated material. 26. Remove the extracts carefully from the Falcon tubes and transfer them to 15- or 50-ml (depending on the volume) sterile Falcon tubes. 27. Snap-freeze the Falcon tubes containing the samples in liquid nitrogen. The lysed frozen cell extracts are stored at −80°C for future use. 28. Thaw the frozen Falcon tubes containing the lysed cell extracts by placing the tubes in cold water. Centrifuge the tubes containing the lysed cell extracts in an Eppendorf centrifuge 5810R using a A-4–62 rotor at 751 × g (4,000 rpm) for 5 min at 4°C to remove any precipitated material. 3.1.2. Protein Purification IgG-Sepharose Beads

1. All purification steps are performed at 4°C with precooled buffers and equipment. 2. The Falcon tubes containing the lysed frozen cell extracts are thawed by placing the tubes in cold water. 3. Prior to use, add 1 ml of IgG-Sepharose beads to a fresh 15-ml Falcon tube. The beads are washed with 3 × 5 ml of IPP buffer. Briefly mix the contents by tilting the Falcon tubes upside down. 4. Centrifuge the Falcon tubes containing the IgG-Sepharose beads with IPP buffer in an Eppendorf centrifuge 5810R using a A-4–62 rotor at 751 × g (4,000 rpm) for 5 min at 4°C. Discard the supernatants, taking care not to disturb the beads. 5. Add 100 ml of the washed IgG-Sepharose beads and 100 ml of IPP buffer to each Falcon tube containing cell extract. Briefly mix the contents by tilting the Falcon tubes upside down.

Systematic Characterization of the Protein Interaction Network

197

6. Rotate the Falcon tubes for 3 h at 4°C using a LabQuake shaker (Barnstead/Thermolyne). 7. Centrifuge the Falcon tubes containing the IgG beads and extracts in an Eppendorf centrifuge 5810R using a A-4–62 rotor at 751 × g (4,000 rpm) for 5 min at 4°C. Remove the supernatants as much and as carefully as possible, taking care not to disturb the loose bead pellets. 8. Resuspend the bead pellets in the leftover supernatants and transfer the beads into 0.8 × 4 cm Bio-Rad polypropylene prep columns (Bio-Rad). Make use of P1000 pipette tips clipped at the ends to transfer the beads into the prep columns. Remove the bottom outlet plugs of the columns and allow the eluates to drain by gravity flow. 9. Wash the columns five times with 200 ml of IPP buffer and twice with 200 ml of TEV cleavage buffer. 10. Close the bottom outlet of the column and add 200 ml of 1× TEV cleavage buffer and 5 ml of TEV protease (2 mg/ ml). Close the top of the column with a cap and rotate the column overnight at 4°C. Calmodulin-Sepharose Beads

1. Remove the top and bottom outlet plugs of the columns after incubation with the TEV protease and recover the eluates by gravity flow into Eppendorf tubes. Wash the columns with 200 ml TEV cleavage buffer and collect the eluates into the same Eppendorf tubes containing the eluates recovered after TEV cleavage. Mix the contents in the Eppendorf tubes by gently pipetting up and down. 2. To the eluates add 400 ml of calmodulin binding buffer and 3 ml of 1 M CaCl2. 3. Transfer the mixtures into 0.8 × 4 cm Bio-Rad polypropylene prep columns containing 200 ml of calmodulin-Sepharose beads (Amersham Biosciences) washed twice with 5 ml of calmodulin binding buffer. 4. Close the tops of the columns with caps and rotate for 2 h at 4°C using a LabQuake shaker. 5. Remove the top and bottom plugs of the columns and drain the eluates by gravity flow. 6. The beads are washed five times with 200 ml of calmodulin binding buffer followed by three washes with 200 ml of calmodulin wash buffer. 7. The bound proteins are eluted in six fractions of 100 ml into fresh Eppendorf tubes using calmodulin elution buffer (see Note 7). 8. The eluted fractions (600 ml) are distributed into two separate Eppendorf tubes in equal volumes of 300 ml.

198

Babu et al.

9. Dry down 300 ml of the eluted fractions using a Speedvac (Eppendorf Vacufuge). Add 60 ml of 2× SDS sample buffer to the dried eluate. The eluate and the sample buffer mixture are boiled for 5 min. The proteins are separated by SDSPAGE and visualized by silver staining. Protein bands excised from the gel are analyzed by MALDI-TOF mass spectrometry. 10. In parallel with step 9, the other 300-ml eluted fraction is dried down using a Speedvac. This dried sample is then subjected to LC-MS/MS for protein identification. 3.2. Protein Identification by MALDI-TOF Mass Spectrometry

3.2.1. SDS-Polyacrylamide Gel Electrophoresis

The tagged and purified S. cerevisiae proteins are separated by SDS-PAGE and stained with silver. The protein bands are excised, reduced, alkylated, subjected to in-gel tryptic digestion, and analyzed by MALDI-TOF mass spectrometry. The mass spectra are searched using Knexus automation against the complete yeast protein database to generate probable identifications of the proteins. The identified proteins are associated back to the purified protein bands (Fig. 1b). The various steps, such as silver staining, in-gel trypsin digestion, extraction and purification of tryptic peptide fragments, spotting samples on MALDI target plates, acquisition of spectra, and protein identification, are described below: 1. The steps described below are carried out using the Whatman Model V16 (Gibco BRL) gel system. The glass plates, spacers, and combs should be clean and free of dried gel fragments, grease, and dust. 2. For each gel, lay out one small and one large glass plate, separated by spacers. 3. Carefully slide the glass plates into the holder by making sure that the glass plates are pushed all the way to the bottom. 4. Slide in the combs and mark lines at 2–3 cm from the bottoms of the combs. 5. Remove the combs gently and pour 12.5% polyacrylamide resolving gels up to the marked lines. Overlay a thin layer of water-saturated n-butanol on the top of each gel. Allow the gels to polymerize for about 30 min. 6. Invert the gels to remove the n-butanol. Touch with filter paper to wick off residual liquid. 7. Pour the 4.5% polyacrylamide stacking gels to fill the remaining space between the glass plates. Insert the combs and allow the gels to polymerize for another 30 min. 8. Once the stacking gels are polymerized, carefully remove the combs and use a 3-ml syringe fitted with a 22G needle to wash the wells with running buffer. 9. Place the gels in the unit and add running buffer to the upper and lower chambers of the unit.

Systematic Characterization of the Protein Interaction Network

199

10. Load each well with 60-ml samples suspended in 2× SDS sample buffer. Include high and low range Precision Plus prestained protein molecular weight standards (Bio-Rad) in two of the wells for each gel. 11. Complete the assembly of the gel unit by attaching the power cords first to the apparatus, and then to the power supply. 12. Turn on the power supply and run the gels at 150 V through the stacking gels and 200 V through the resolving gels. The gels are run until the blue dye front reaches the bottom. 13. After the electrophoresis, disassemble the gel plate from the apparatus. Use a thin spatula to carefully pry the upper glass plates away from the gels. 14. Transfer the gels immediately to clean the staining solution containing fixer and proceed immediately with the silverstaining protocol described below. 3.2.2. Silver-Staining the SDS-Polyacrylamide Gel

1. Agitate the gel in clean fixer for 20 min on a rocking shaker. 2. Rinse the gel with 20% ethanol for 10 min. 3. The gel is then washed twice with 500 ml double-distilled water for 10 min. Note that thorough rinsing gives a uniform, low background. 4. Remove the double-distilled water and agitate the gel in 500 ml sensitizer solution for 1 min. 5. Remove the sensitizer solution and rinse the gel twice again with double-distilled water for 20 s. 6. Pour off the water and incubate the gel in 200 ml of 0.1% silver nitrate for 30 min. 7. Discard the silver nitrate and rinse the gel once with distilled water for 20 s to remove excess silver nitrate. 8. Wash the gel with 50–75 ml of freshly prepared developing solution for half a minute. Replace with fresh developing solution and agitate the gel slowly by hand constantly (see Note 8). 9. When the desired staining intensity is achieved, discard the developing solution and add 80 ml of stop solution to the gel. 10. Incubate the gel with the stop solution for a minimum of 20 min before proceeding to excise the protein bands from the silver-stained gels.

3.2.3. Protein Identification by Gel-Based MALDI-TOF Mass Spectrometry Reduction and Alkylation of Protein Bands

1. Protein bands are excised from silver-stained gels with clean razor blades. The gel slices are excised as closely as possible to the boundaries of the protein bands and stored in −80°C freezers in 96-well polypropylene plates (Nunc). 2. Gel slices in 96-well polypropylene plates are thawed for approximately 20 min. The liquid that accumulates during thawing is carefully removed.

200

Babu et al.

3. Gel slices are shrunk with 200 ml of 100% ACN for ~10 min on an IKA Schuttler MTS 4 orbital shaker (VWR Scientific) at 700 rpm (see Note 9). Remove all liquid. 4. The gel slices are reduced with 75 ml of 100 mM NH4HCO3 containing 10 mM DTT for 30 min in a 50°C heating block. Note that the gel pieces should be covered with liquid. When hydrated, there should still be some fluid left. 5. Remove all liquid by centrifuging the plate at 58 × g (500 rpm) in a Beckman Allegra X-12 centrifuge using a SX4750 m plate carrier for 3–5 min, and repeat step 3. 6. The gel slices are alkylated in the 96-well microtiter plate with 75 ml of 100 mM NH4HCO3 containing 55 mM iodoacetamide for 20 min in the dark at room temperature. 7. Centrifuge the plate at low speed 58 × g (500 rpm) for 3–5 min. Remove all liquid. Repeat step 3. In-Gel Tryptic Digestion

1. The gel slices are hydrated with 60 ml of digestion buffer containing trypsin for 30–45 min on ice. Note that the gel pieces should be fully hydrated with trypsin solution on ice. After 15–20 min, if needed, add more digestion buffer containing trypsin to allow complete hydration. 2. Add 20 ml (if needed) of digestion buffer without trypsin and incubate samples overnight at 37°C (see Note 10).

Extraction of Tryptic Peptides

1. The extracted peptides are transferred into clean 96-well polypropylene plates. 2. Add 100 ml of 100 mM NH4HCO3 to the gel slices and extract peptides by shaking at 700 rpm on an orbital shaker for 60 min at room temperature. 3. Briefly centrifuge, transfer the extracted peptides into the clean 96-well polypropylene plates, and add 2.5 ml of 100% AA to the extracted peptides in each well.

Purification of Tryptic Peptides

1. Purification is performed using bulk C18 reverse phase resin (Sigma). Add 1.5 g of dry resin to a reservoir. 2. Wash the dry resin two times with HPLC-grade methanol and two times with HPLC-grade 66% ACN, 1% AA prior to use. Add 75% ACN, 1% AA to prepare 5:1 resin slurry. 3. Add 2.5 ml of C18 reverse phase resin slurry to the extracted peptides in the wells of the 96-well plate. The resin should float on top of the liquid. Shake the plate on an orbital shaker at 500–700 rpm for 45 min at room temperature. Discard the liquid underneath the beads by using a 200-ml multichannel pipette. 4. Add 200 ml of 2% ACN, 1% AA. Shake the plate for 5–15 min on an orbital shaker at 500–700 rpm at room temperature.

Systematic Characterization of the Protein Interaction Network

201

5. Prepare in advance a 384-well Melt Blown Polypropylene (MBPP) Whatman filter plate (Whatman). Place the MBPP filter plate on top of a 384-well collection plate (Whatman). Wash the MBPP filter plate wells with 15 ml of 66% ACN, 1% AA. Centrifuge the MBPP filter plate for 1–2 min at 1,000g and discard the filtrates collected in the collection plate. 6. Centrifuge the 96-well plates from step 4 for 5–15 min at 1,000g at room temperature. Remove the supernatants with a 200-ml multichannel pipette and discard them. 7. Elute the peptides by adding 30 ml of 66% ACN, 1% AA. Shake the 96-well plate briefly at high speed on an orbital shaker. Ensure that all of the resin has entered into the slurry. Incubate the plates for approximately 5–10 min at room temperature. The resin should make a slurry and slowly pellet to the bottom. 8. Transfer the liquid supernatant with 50-ml multichannel pipette from the 96-well plates to the 384-well MBPP filter plate. Place a 384-well collection plate (Whatman) under the MBPP filter plate and centrifuge for 3–5 min at 2,000g. The filtrates collected in the 384-well collection plate are either spotted immediately onto the Bruker MALDI target plate or sealed with sealing foil and stored at −70°C until spotting onto MALDI targets. Spotting Samples onto the MALDI Target Plates

1. Separate the Bruker MALDI target plate from its base. 2. Remove the previously spotted matrix spots from the MALDI target plate by washing the plate with 100% methanol followed by rinsing the plates gently with HPLC-grade water. Wipe the plate with 100% methanol using Kimwipes (see Note 11). 3. Spot 1 ml of a-cyano-4-hydroxycinnamic acid matrix (Fluka Buchs SG, Switzerland) solution and 1 ml of a purified trypsindigested sample onto each target spot of the MALDI target plate. Allow the samples to dry at room temperature. 4. In the bottom-most row of the MALDI target plate, spot 1 ml of a-cyano-4-hydroxycinnamic acid matrix solution and 1 ml of peptide calibration standard (Bruker). 5. After the spotting is completed, reseal the 384-well trypsindigested peptide plate (s) and store at −80°C. Make sure that all the spots on the MALDI target plate are dry before acquiring the spectra.

Acquiring Spectra from MALDI Target Plates

1. Double click on the flex control icon of an ULTRAFlex II MALDI-TOF instrument. When the log on information appears, Click OKAY. Open the FLEX control method: RP_ pepmix.par. Wait for the method to upload.

202

Babu et al.

2. Insert the spotted MALDI target plate into the source chamber and click on the green load button on the instrument. 3. Wait until the system fully loads the MALDI plate. Once the plate is loaded, the system status light on the FLEX control should turn green and say “READY.” The sample status light should turn green and say “IN.” 4. The spectrum is acquired by clicking the start button and adjusting the laser intensity so that the peptide peaks can be visualized. 5. Click on “Add” to the sum buffer. 6. Click on the “calibration tab” and choose “peptidemix monoisotopic II” from the drop down menu, and a peptide reference list will appear. 7. Change the zooming factor to 0.5. 8. To manually calibrate the instrument, click on each peptide in the reference list. The red vertical line indicates the corresponding peptide. Click on the monoisotopic peak (left of the highest peak) to visualize each of the peptides. 9. Click on “accept fit result”; now the instrument is calibrated. 10. To automatically acquire the samples, click on the AutoXecute tab. Load the.txt file by clicking on the select button. Make sure that the method to be run is “autoXYMP.” Check the box where it shows the autoX output. 11. Click on “START automatic run” and the system should automatically start acquiring the spectra. MALDI-TOF Spectral Analysis and Protein Identification

1. MALDI spectral searches are performed using Knexus automation, a Windows-based program from Genomics Solutions Bioinformatics (Discovery Scientific, Inc., Vancouver, Canada) that selects the spectral peaks automatically and performs the searches. 2. After the installation of the Knexus program, the most recent Fasta sequences of the S. cerevisiae genome downloaded from the European Bioinformatics Institute database (http://www.ebi.ac.uk/) were uploaded into the machine using the Knexus Database Installation Wizard. 3. Protein identification is done using the ProFound search engine, which matches the observed peaks against the database of theoretical peaks. 4. The program runs ProFound on all the spectra using one set of conditions. In addition, a Java program was developed in-house, which automates the rerunning of Knexus. It uses 72 varying parameter sets and evaluates the aggregate results to produce a set of identified proteins in each case. Based on these results, it calculates an aggregate score for each protein.

Systematic Characterization of the Protein Interaction Network

203

5. Using graphical interface software developed in-house, the user is able to specify where the bands are located on the corresponding gel image. The entire area of the band is inputted so that its total intensity is measured (see Note 12). When all the data has been entered into the system for a given band, an annotated gel image and a plot are produced in JPEG format. Examples of affinity-purified protein complexes identified by in-gel trypsin digestion and MALDI mass spectral analysis are shown in Fig.2.

a Cop1-TAP

b Sec27-TAP Sec27/Cop1

Cop1/Sec27

Cop1*

Cop1/Sec27

Sec27/Cop1

Cop1/Sec27/Sec26 Cop1/Sec27/Sec26 Ret2/Cop1 Eno2/Sec21

Sec27/Sec26 Sec27/Sec21

Sec26 Sec27/Sec21

Ret2

Pho85/Sec21 Sec26/Cop1

Sec21/Sec27 Pho85/Elp3 Pho80

Sec27/Sec26 Ydj1

Cop1 Eno2/Sec26

Cop1 Sec27*

Sec27/Sec26/Cop1/Sec21 Sec21/Sec27/Sec26

Sec27/Cop1

Sec21

Mre11

Ret2/Pfk26/Ded1 Cdc19

Sec27

Sec28

Sec28

Sec28

Sec28

Ret3 Ret3

Fig. 2. Examples of silver-stained SDS polyacrylamide gel tracks containing COP1 coatomer protein complexes, which are involved in ER-to-Golgi and Golgi-to-ER transport, following affinity purification of S. cerevisiae TAP-tagged Cop1 (a) and Sec27 (b) proteins, respectively. Purifications of TAP-tagged Cop1 and Sec27 revealed other subunits of the COP1 coatomer protein complexes. The TAP-tagged bait proteins are specified at the top of each lane. Individual subunits of the purified complex and various co-purifying contaminants, all indicated by arrowheads, were identified by in-gel trypsin digestion followed by MALDI-TOF mass spectrometry. Asterisks designate the intact tagged bait protein recovered from each purification; proteolytic fragments of various coatomer subunits are also often identified. The known coatomer subunits include Cop1, Sec21, Sec26, Sec27, Sec28, Ret2, and Ret3. These purified preparations of coatomer complex also contained small amounts of the contaminating proteins Eno2, Pho80, Pho85, Elp3, Ded1, Mre11, Ydj1, Pfk26, and Cdc19.

204

Babu et al.

3.3. Protein Identification by Gel-Free Liquid ChromatographyTandem Mass Spectrometry 3.3.1. Proteolysis and Sample Preparation for LC-MS/MS

1. Three hundred ml eluted fractions (described in step 10 in “Calmodulin-Sepharose Beads”) in microcentrifuge tubes were dried down using a Speedvac. 2. The dried down samples are dissolved in 20 ml of the immobilized trypsin solution. Make sure to dissolve all the dried down sample in the immobilized trypsin solution. Using a P1000 pipette tip, pipette the samples up and down. 3. The samples are incubated overnight at 30°C with rotation or agitation. 4. Add 20 ml of the LC-MS solvent A to stop digestion. 5. The peptide mixtures are centrifuged for 5 min at maximum speed using an Eppendorf centrifuge. 6. Carefully transfer 10–20 ml of the supernatants containing the peptide mixtures into fresh microcentrifuge tubes and immediately analyze by LC-MS, or store the samples at −20°C prior to use.

3.3.2. LC-MS Spectral Analysis and Protein Identification

1. Protein samples are analyzed using single-dimension reverse phase chromatography coupled online to ion trap tandem mass spectrometry using standard conditions. 2. The microcolumns are packed with ~10 cm of 5-mm Zorbax eclipse XDB-C18 resin and are interfaced to a custom electrospray ion source. 3. The packed microcolumn is placed in line with the LC-MS instrument. 4. A Proxeon nano HPLC pump is used to deliver a stable tip flow rate of ~300 nl/min during the peptide separations. 5. Peptide elution is achieved using the following gradient which may be increased or decreased according to the complexity of the sample: 15% of solvent B from 0 to 30 min, 40% of solvent B from 31 to 49 min, 80% of solvent B from 50 to 55 min, 80% of solvent B from 56 to 60 min, 100% of solvent A from 61 to 65 min, 45% solvent B from 66 to 70 min, 80% solvent B from 71 to 75 min, 80% solvent B from 76 to 80 min, 100% solvent A from 81 to 85 min, 100% solvent A from 86 to 105 min. The flow rate at the tip of the needle is set to 300 nl/min for 105 min. 6. The mass spectrometer cycles run through successive series of 11 scans as the gradient progresses. The first scan in a series is a full mass scan and is followed by successive tandem mass scans of the most intense ions. 7. Proteins from the mixture are identified using the SEQUEST computer algorithm (23) and validated using the STATQUEST (24) probabilistic scoring program (see Note 13).

Systematic Characterization of the Protein Interaction Network

205

4. Notes 1. A similar TAP strategy is being used to purify the membraneassociated proteins except that all the purification steps are performed in the presence of a nonionic detergent. 2. Although low-abundance proteins can be detected using highly sensitive mass spectrometers, it is important to start with enough cells for the purification. This can be achieved by noting the expression level of the tagged protein (3) or by examining the literature to identify conditions that are suitable for the preparation of cell extracts specifically for the target protein. It is possible in some cases to successfully purify and identify by mass spectrometry a tagged protein that is not detectable by Western blotting. 3. If using the same coffee grinder to grind multiple protein samples, make sure not to clean the grinder with water because it will freeze. Use wipes and paper towels to clean the grinder. 4. Make sure to mix the YEB buffer with the lysed powder until it thaws. The tube is then inverted several times at room temperature to ensure proper mixing. 5. In some cases, however, it is necessary to use a higher salt concentration to reduce nonspecific binding of proteins. 6. It is not mandatory to perform the dialysis step in the preparation of extracts. 7. Make sure to elute the samples with calmodulin elution buffer, as it is a very critical step in releasing the complexes from the resin. Yields of the bound proteins can sometimes be improved either by increasing the salt concentration or by increasing the EGTA concentration in the calmodulin elution buffer. 8. When the developing solution becomes cloudy, replace it with new solution. Make sure to develop all the gels for about the same time. 9. The gel slices will become white and should feel gritty like grains of sand. Make sure to remove all the liquid when the gel slices are completely shrunk. 10. Make sure that in each well there is enough liquid for the gel slices to be completely submerged. 11. If the spots still remain on the MALDI target plate, sonicate the plate for 10 min with 100% methanol followed by 5 min sonication with HPLC water and 100% methanol, respectively.

206

Babu et al.

12. The band location algorithm tries to identify the complete area of each band. Make sure to confirm that the location of the band that the computer identifies matches the band in the gel image. 13. The acquired spectra are used to search the S. cerevisiae protein database downloaded from the European Bioinformatics Institute (http://www.ebi.ac.uk/) using the SEQUEST search algorithm (23) to identify the proteins from which they originated. Confidence scores for all putative matches were evaluated and assigned using a probabilistic STATQUEST (24) scoring program. Proteins detected with two or more high-confidence peptide matches and with a minimum likelihood threshold cut-off of 90% or greater probability were considered as a positive identification.

Acknowledgments The authors wish to thank Xinghua Guo and Gouqing Zhong for purifying the S. cerevisiae bait proteins and Shamanta Chandran, Peter Wong, and Constantine Christopoulos for assisting with mass spectrometry. This research was supported by grants from Genome Canada and the Ontario Genomics Institute to J.F.G. and A.E. References 1. Puig, O., Caspary, F., Rigaut, G., Rutz, B., Bouveret, E., Bragado-Nilsson, E., Wilm, M., and Seraphin, B. (2001) The tandem affinity purification (TAP) method: A general procedure of protein complex purification. Methods 24, 218–229. 2. Rigaut, G., Shevchenko, A., Rutz, B., Wilm, M., Mann, M., and Seraphin, B. (1999) A generic protein purification method for protein complex characterization and proteome exploration. Nat. Biotech. 17, 1030–1032. 3. Ghaemmaghami, S., Huh, W.-K., Bower, K., Howson, R.W., Belle, A., Dephoure, N., O’Shea, E.K., and Weissman, J.S. (2003) Global analysis of protein expression in yeast. Nature 425, 737–741. 4. Gavin, A.C., Bosche, M., Krause, R., Grandi, P., Marzioch, M., Bauer, A., Schultz, J., Rick, J.M., Michon, A.M., Cruciat, C.M., Remor, M., Hofert, C., Schelder, M., Brajenovic, M., Ruffner, H., Merino, A., Hudak, M., Dickson, D., Rudi, T., Gnau, V., Bauch, A., Bastuck, S.,

Huhse, B., Leutwein, C., Heurtier, M.A., Copley, R.R., Edelmann, A., Querfurth, E., Rybin, V., Drewes, G., Raida, M., Bouwmeester, T., Bork, P., Seraphin, B., Kuster, B., Neubauer, G., and Superti-Furga, G. (2002) Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 415, 141–147. 5. Gavin, A.C., Aloy, P., Grandi, P., Krause, R., Boesche, M., Marzioch, M., Rau, C., Jensen, L.J., Bastuck, S., Dumpelfeld, B., Edelmann, A., Heurtier, M.A., Hoffman, V., Hoefert, C., Klein, K., Hudak, M., Michon, A.M., Schelder, M., Schirle, M., Remor, M., Rudi, T., Hooper, S., Bauer, A., Bouwmeester, T., Casari, G., Drewes, G., Neubauer, G., Rick, J.M., Kuster, B., Bork, P., Russell, R.B., and Superti-Furga, G. (2006) Proteome survey reveals modularity of the yeast cell machinery. Nature 440, 631–636. 6. Krogan, N.J., Cagney, G., Yu, H., Zhong, G., Guo, X., Ignatchenko, A., Li, J., Pu, S., Datta, N., Tikuisis, A.P., Punna, T., Peregrin-Alvarez,

Systematic Characterization of the Protein Interaction Network

7.

8.

9.

10.

11.

12.

13.

J.M., Shales, M., Zhang, X., Davey, M., Robinson, M.D., Paccanaro, A., Bray, J.E., Sheung, A., Beattie, B., Richards, D.P., Canadien, V., Lalev, A., Mena, F., Wong, P., Starostine, A., Canete, M.M., Vlasblom, J., Wu, S., Orsi, C., Collins, S.R., Chandran, S., Haw, R., Rilstone, J.J., Gandi, K., Thompson, N.J., Musso, G., St Onge, P., Ghanny, S., Lam, M.H., Butland, G., Altaf-Ul, A.M., Kanaya, S., Shilatifard, A., O’Shea, E.,Weissman, J.S., Ingles, C.J., Hughes, T.R., Parkinson, J., Gerstein, M., Wodak, S.J., Emili, A., and Greenblatt, J.F. (2006) Global landscape of protein complexes in the yeast Saccharomyces cerevisiae. Nature 440, 637–643. Shoemaker, B.A., and Panchenko, A.R. (2007) Deciphering protein-protein interactions. Part I. Experimental Techniques and Databases. PLoS Computational Biol. 3, 0337–0344. Uetz, P., Giot, L., Cagney, G., Mansfield, T.A., Judson, R.S., Knight, J.R., Lockshon, D., Narayan, V., Srinivasan, M., Pochart, P., Qureshi-Emili, A., Li, Y., Godwin, B., Conover, D., Kalbfleisch, T., Vijayadamodar, G., Yang, M., Johnston, M., Fields, S., and Rothberg, J.M. (2000) A comprehensive analysis of protein–protein interactions in Saccharomyces cerevisiae. Nature 403, 623–627. Ito, T., Chiba, T., Ozawa, R., Yoshida, M., Hattori, M., and Sakaki, Y. (2001) A comprehensive two-hybrid analysis to explore the yeast protein interactome. Proc. Natl. Acad. Sci. U S A 98, 4569–4574. Polanowska, J., Martin, J.S., Fisher, R., Scopa, T., Rae, I., and Boulton, S.J. (2004) Tandem immunoaffinity purification of protein complexes from Caenorhabditis elegans. Biotechniques 36, 778–780. Veraksa, A., Bauer, A., and Tsakonas, S.A. (2005) Analyzing protein complexes in Drosophila with tandem affinity purification-mass spectrometry. Dev. Dyn 232, 827–834. Butland, G., Peregrín-Alvarez, J.M., Li, J., Yang, W., Yang, X., Canadien, V., Starostine, A., Richards, D., Beattie, B., Krogan, N., Davey, M., Parkinson, J., Greenblatt, J., and Emili, A. (2005) Interaction network containing conserved and essential protein complexes in Escherichia coli. Nature 433, 531–537. Tsai, A., and Carstens, R.P. (2006) An optimized protocol for protein purification in cultured mammalian cells using a tandem affinity

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

207

purification approach. Nat. Protoc. 1, 2820– 2827. Knuesel, M., Wan, Y., Xiao, Z., Holinger, E., Lowe, N., Wang, W., and Liu, X. (2003) Identification of novel protein-protein interactions using a versatile mammalian tandem affinity purification expression system. Mol. Cell Proteomics 2, 1225–1233. Brown, A.P., Affleck, V., Fawcett, T., and Slabas, A.R. (2006) Tandem affinity purification tagging of fatty acid biosynthetic enzymes in Synechocystis sp. PCC6803 and Arabidopsis thaliana. J. Exp. Bot. 57, 1563–1571. Oliver, S.G. (2000) Guilt by association goes global. Nature 403, 601–603. Collins, S.R., Kemmeren, P., Zhao, X.C., Greenblatt, J.F., Spencer, F., Holstege, F.C., Weissman, J.S., and Krogan, N.J. (2007) Towards a comprehensive atlas of the physical interactome of Saccharomyces cerevisiae. Mol. Cell. Proteomics 6, 439–450. Pu, S., Vasblom, J., Emili, A., Greenblatt, J., and Wodak, S.J. (2007) Identifying functional modules in the physical interactome of Saccharomyces cerevisiae. Proteomics 7, 944–960. Grisshammer, R., and Tate, C.G. (1995) Overexpression of integral membrane proteins for structural studies. Q. Rev. Biophys. 28, 315–422. Loll, P.J. (2003) Membrane protein structural biology: the high throughput challenge. J. Struct. Biol. 142, 144–153. Tate, C.G., and Grisshammer, R. (1996) Heterologous expression of G-protein-coupled receptors. Trends Biotechnol. 14, 426–430. Dobrovetsky, E., Lu, M.L., Broza, R.A., Khutoreskaya, G., Bray, J.E., Savchenko, A., Arrowsmith, C.H., Edwards, A.M., and Koth, C.M. (2005) High-throughput production of prokaryotic membrane proteins. J. Struct. Funct. Genomics 6, 33–50. Eng, J.K., McCormack, A.L., and Yates, J.R. III (1994) An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J. Am. Soc. Mass Spectrom. 5, 976–989. Kislinger, T., Rahman, K., Radulovic, D., Cox, B., Rossant, J., and Emili, A. (2003) PRISM, a generic large scale proteomic investigation strategy for mammals. Mol. Cell Proteomics 2, 96–106.

Chapter 12 Protein Microarrays Joseph Fasolo and Michael Snyder Summary Protein microarrays containing nearly the entire yeast proteome have been constructed. They are typically prepared by overexpression and high-throughput purification and printing onto microscope slides. The arrays can be used to screen nearly the entire proteome in an unbiased fashion and have enormous utility for a variety of applications. These include protein–protein interactions, identification of novel lipid- and nucleic acid-binding proteins, and finding targets of small molecules, protein kinases, and other modification enzymes. Protein microarrays are thus powerful tools for individual studies as well as systematic characterization of proteins and their biochemical activities and regulation. Key words: Protein chips, Microarrays, High throughput, Protein interactions

1. Introduction Protein microarrays contain a large number of proteins that have been spotted in an addressable format and at high density onto microscope slides, thereby allowing the individual analysis of large numbers of proteins simultaneously (1, 2). There are two types of protein microarrays: antibody microarrays that are used for protein profiling, and functional protein microarrays that contain sets of proteins or, for the case of yeast, nearly an entire yeast proteome. The functional arrays are used for a wide variety of applications and are presented here. Protein arrays are typically prepared from large overexpression libraries in which proteins are overexpressed, purified in a 96-well format, and then spotted at high density onto microscope slides. The arrays are then probed using molecules containing fluorescent or other probes. To date, protein microarrays have been prepared I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_12, © Humana Press, a part of Springer Science + Business Media, LLC 2009

209

210

Fasolo and Snyder

for yeast (6,000), Arabidopsis (5,000), humans (8,000), and coronaviruses, as well as a number of bacteria (3–7). Thus far, protein microarrays have been used for a wide variety of applications. These include enzymatic assays and interactions with proteins, lipids, small molecules, and nucleic acids (3, 8). One advantage of protein microarrays is that an entire proteome can be screened in an unbiased fashion, thereby leading to the discovery of novel activities that were not anticipated. For example, screening of a yeast protein chip using biotinylated liposomes revealed many new lipid-binding proteins including enzymes involved in glycolysis (3), and screening of a yeast proteome chip with yeast DNA revealed a novel metabolic enzyme, Arg5,6, that is associated with DNA and appears to regulate mitochondrial gene expression (9). Protein microarrays have been used for many other applications as well. Recently, arrays that contain most yeast transcription factors have been prepared and probed using motifs that are conserved across yeast species but whose binding protein was not known (10). This study assigned candidate binding proteins for specific sequences, one of which was confirmed. Protein microarrays have also been used to assess antibody specificity (11) and for identification of autoreactive antibody in patients with diseases such as autoimmune diseases and cancer and thereby find candidate biomarkers and insights into the disease state (5). Lastly, protein microarrays can be used to identify substrates of modification enzymes such as kinases and ubiquitation enzymes (12, 13).

2. Materials 2.1. Generation of C-Terminally Tagged Yeast Strains

1. pDONR221 vector containing the open reading frame (ORF) of interest. 2. pBG1805 vector: c-terminal His6X-3C-Protein A “zz” domain or pYES-DEST52 for His6X-V5 fusion proteins. 3. Flanking sequences for integration into Gateway pDONR221 plasmid (a) att B4: 5¢ GGGGCAACTTTGTATAGAAAAGTTG 3¢ (added to the 5¢ PCR primer). (b) att B1: 5¢ GGGGCTGCTTTTTTGTACAAACTTG 3¢ (added to the 3¢ PCR primer). 4. Sequencing primers for pBG1805 plasmid (a) F5: 5¢ CATTTTCGGTTTGTATTACTTCTTATTC 3¢. (b) R3: 5¢ GGACCTTGAAAAAGAACTTC 3¢. 5. S. cerevisiae genomic DNA (strain BY4700, MATa ura3D0).

Protein Microarrays

211

6. S. cerevisiae Y258: MATa, pep4, his4-58, ura3-52, leu 2-3, 112. 7. Pfx polymerase (Invitrogen) or Pfu Ultra polymerase (Stratagene). 8. BsrG1 endonuclease for screening positive clones. 9. Culture tubes. 10. Agarose. 11. LB plus Kanamycin, both liquid, and 2% bacto-agar. (a) 10 g tryptone. (b) 5 g yeast extract. (c) 5 g NaCl. (d) 20 g agar (for plates). (e) Make up to 1 L with distilled water (dH2O). (f) Autoclave and allow to cool prior to addition of 50 mg/ mL of ampicillin final concentration. 12. DNA ladder 13. E. coli: DH5a 14. TAE (a) 242 g Tris base. (b) 57.1 mL 100% glacial acetic acid. (c) 100 mL 0.5 M ethylenediaminetetraacetic acid (EDTA) (pH 8.0). (d) Make up with dH2O up to 1 L. 15. LR clonase (Invitrogen). 16. BP clonase (Invitrogen). 2.2. Transformation Reagents

1. YPAD (Yeast extract-peptone-adenine-dextrose) (a) 10 g bacto yeast extract. (b) 20 g bacto peptone. (c) 50 mg adenine. (d) 20 g dextrose. (e) 20 g agar (for plates). (f) Make up with dH2O up to 1 L. 2. 1 M lithium acetate (LiAc). 3. 50% polyethylene glycol (PEG). 4. 10 mg/mL ssDNA. 5. DNA. 6. dH2O. 7. Water bath or heat block. 8. 96-well box.

212

Fasolo and Snyder

9. 96-well polymerase chain reaction (PCR) plate. 10. Synthetic complete minus uracil media (Sc-ura) (a) 1.5 g yeast nitrogen base. (b) 5 g (NH4)2SO4. (c) 2 g Sc-ura drop out mix (commercially available). (d) 20 g raffinose (or dextrose for starter culture media and SD-ura plates). (e) 20 g agar (for plates). 2.3. High-Throughput Immunoblot Analysis

1. 3× YEP–GAL (yeast extract–peptone–galactose) (a) 30 g yeast extract. (b) 60 g peptone. (c) Make up to 700 mL with dH2O. (d) Add 300 mL of sterile filtered 20% galactose to media after autoclaving. 2. 0.5 mm zirconia beads or acid-washed glass beads. 3. Lysis buffer 150 (a) 50 mM Tris–HCL at pH 7.5. (b) 150 mM NaCl. (c) 1 mM ethyleneglycoltetraacetic acid (EGTA). (d) 10% glycerol. (e) 0.1% Triton X-100. (f) 0.5 mM dithiothreitol (DTT). (g) 1 mM phenylmethylsulfonyl fluoride (PMSF). (h) 1× complete protease inhibitor tablet (Roche). 4. Paint shaker (5G-HD, Harbil) or similar mechanism to break cells. 5. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) gels. 6. Polyvinylidene fluoride (PVDF) or nitrocellulose membranes. 7. Multichannel pipette. 8. Semi-dry transfer apparatus. 9. HRP-IgG. 10. HA antibody (16B12, Covance). 11. Imaging Film (Kodak BioMax MR Film).

2.4. Preparation of Proteins for Protein Microarray

1. Wash buffer (150) (a) 50 mM Tris-HCl (pH 7.5). (b) 150 mM NaCl. (c) 10% glycerol. (d) 0.1% and Triton X-100.

Protein Microarrays

213

2. Elution buffer (a) 50 mM Tris (pH 7.5). (b 150 mM NaCl. (c) 25% glycerol. (d) 0.1% Triton X-100. 3. 96-well box. 4. PVDF 96-well filter plate (1.2-mm pore size). 5. IgG-Fast Flow 6 Sepharose. 6. 3C protease. 7. 384-well microplate (to array proteins). 8. Adhesive foil lids for sealing microplates. 2.5. Printing Arrays

1. Arrays suitable for assay being performed. 2. Bio-Rad ChipWriter Pro or comparable printer.

2.6. Purifying V5-Fusion Protein Probe

1. Phosphate buffered saline (PBS) lysis solution (a) PBS (0.01 M phosphate buffer, 0.0027 M potassium chloride, 0.137 M sodium chloride; pH 7.4). (b) 0.1% Triton X-100. (c) 0.5 mM DTT. (d) 2 mM MgCl2. (e) 500 mM NaCl. (f) 50 mM imidazole (pH 7.4). (g) 1 mM PMSF. (h) Complete protease inhibitor cocktail – EDTA-free (Roche). (i) Phosphatase inhibitor Cocktail 1 (Sigma). 2. Ni2+, or Co2+ affinity resin. 3. G25 columns. 4. Microcentrifuge. 5. Wash buffer (a) PBS (0.01 M phosphate buffer, 0.0027 M potassium chloride, 0.137 M sodium chloride; pH 7.4). (b) 0.1% Triton X-100. (c) 500 mM NaCl. (d) 0.5 mM DTT. (e) 50 mM imidazole (pH 7.4). (f) 2 mM MgCl2. 6. Elution buffer (a) PBS (0.01 M phosphate buffer, 0.0027 M potassium chloride, 0.137 M sodium chloride; pH = 7.4).

214

Fasolo and Snyder

(b) 0.1% Triton X-100. (c) 500 mM NaCl. (d) 0.5 mM DTT. (e) 500 mM imidazole (pH 7.4). (f) 2 mM MgCl2. (g) 20% glycerol (if freezing at −80°). 2.7. Probing Arrays

1. Blocking buffer (a) PBS (0.01 M phosphate buffer, 0.0027 M potassium chloride, 0.137 M sodium chloride; pH 7.4). (b) 1% bovine serum albumin (BSA). (c) 0.1% Tween-20. 2. Probe buffer (a) PBS (0.01 M phosphate buffer, 0.0027 M potassium chloride, 0.137 M sodium chloride; pH 7.4). (b) 2 mM MgCl2. (c) 0.5 mM DTT. (d) 0.05% Triton X-100. (e) 50 mM NaCl. (f) 500 µM ATP (for kinases). (g) 1% BSA.

2.8. Identifying Kinase–Substrate Interactions

1. Lysis Buffer (a) 100 mM Tris–HCl pH 7.4. (b) 100 mM NaCl, 1 mM EGTA. (c) 0.1% 2-mercaptoethanol. (d) 0.1% Triton X-100, protease cocktail (Roche). (e) 1 mM EDTA. (f) 50 mM NaF. (g) 10 mM sodium glycerophosphate. (h) 1 mM Na3VO4. 2. Kinase Buffer (a) 100 mM Tris–HCl pH 8.0. (b) 100 mM NaCl. (c) 10 mM MgCl2. (d) 20 mM glutathione. (e) 20% glycerol. 3. Superblock (Pierce) 4. [33P]ATP

Protein Microarrays

215

3. Methods 3.1. High-Throughput Preparation of Clones

Protein microarrays are typically constructed using overexpression libraries. To date, two comprehensive expression libraries have been prepared that are suitable for production of yeast protein microarrays. Initially, a library was prepared in which 5,800 yeast ORFs were fused at their amino terminal coding sequences to glutathione-S-transferase and HisX6 coding sequences (3). Subsequently, a movable ORF, or MORF, library was prepared in which 5,700 yeast ORFs were fused at their carboxy terminal coding sequences to His6 and IgG binding domain of protein A (14). The MORF library contains GATEWAY-compatible sequences flanking the inserts, so the yeast ORFs can be shuttled into any vector. In each library the fusion protein is expressed from the GAL1 galactose promoter. A protocol for library construction is as follows:

3.1.1. Constructing a Large Number of Clones for Protein Arrays

1. Design forward primers containing the att B4 sequence and reverse primers with the att B1 sequence (see subheading 2.1 and Note 4) for the gene of interest. 2. Amplify the gene of interest by PCR using the previously mentioned genomic DNA and polymerase (see subheading 2.1). 3. Perform BP clonase reaction (5 µL) using 30–300 ng of PCR product added to 150 ng of pDONR221 plasmid and following manufacturer’s instructions. 4. Incubate the reaction overnight at 25°C. 5. Combine half of the of the BP reaction (3 µL) with 150 ng of pBG1805 DNA or pYES-DEST52 (for V5 fusion) backbone, 0.6 µL of LR clonase, and 0.6 µL of LR clonase 10× buffer. 6. Incubate the reaction overnight at 25°C and use half of the reaction to transform DH5a cells. 7. Plate the cells on LB/ampicillin (see Subheading 2.1, item 11) plates overnight at 37°C. 8. Miniprep overnight cultures of several colonies of each transformant grown in LB/ampicillin liquid media. 9. Digest with 5 mL of DNA with BsrG1 overnight at 37°C to confirm the presence of insert and sequence clones that contain the insert using sequencing primers (see Subheading 2.1, item 4).

3.1.2. Transforming Yeast with C-Terminal Tagged ORF

1. Inoculate 20 mL YPAD (see subheading 2.2, item 1) with a starter culture of a fresh Y258 colony from YPAD agarose plate. 2. Inoculate 50 mL of YPAD/10 transformations with enough starter culture to attain an OD600 of 0.1.

216

Fasolo and Snyder

3. Grow for several doubling times to an OD600 of ~0.6–0.8. 4. Transfer the cells to a 50-mL conical tube and spin in a tabletop centrifuge at 3,000 rpm/5 min/4°C. 5. Bring up to a volume of 50 mL with water to wash, and repeat centrifugation. 6. Decant water and add 0.5 mL 100 mM LiAc and transfer to a 1.5-mL snap-cap tube. 7. Spin at 13,000 rpm in a microfuge for 30 s and remove the solution with a pipette. 8. Add enough 100 mM LiAc to bring the volume up to 0.5 mL, and vortex to suspend the cells. 9. Aliquot 50 mL of the suspension into separate snap-cap tubes and centrifuge. 10. Remove liquid as before and prepare pellet for transformation. 11. Add 240 µL of 50% PEG solution, 36 mL of 1 M LiAC, 10 mL of 10 mg/mL ssDNA, 2 mL of miniprepped DNA fusion construct, and 72 mL of dH20. 12. Vortex the mixture and incubate at 30°C for 30 min. 13. Next, move the tubes to a water bath set to 42°C for 30 min. 14. Finally, spin the tubes at 10,000 ´ g for 15 s, and aspirate the solution with a pipette. 15. Add 500 mL of dH2O to each tube and vortex briefly. 16. Pipette 50–100 mL of suspension onto Sc-ura/2% dextrose agar Petri dishes (see subheading 2.2, item 10). 17. Allow 48 h to recover and screen the transformants by immunoblot. 3.1.3. High-Throughput Assaying and Preparation of Protein Methods for Assay Protein Expression and High-Throughput Preparation of Proteins High-Throughput Immunoblot Analysis

1. Inoculate 96-well starter cultures in 0.8 mL of SD-ura/2% dextrose. 2. On day 2, wash the cells with Sc-ura/2% raffinose (see subheading 2.2, item 10), and inoculate 5 mL of starter culture into a new 96-well box (2 mL/well) containing 0.8 mL of Sc-ura/2% raffinose with a 3.5-mm glass ball (PGC scientific) in each well for mixing and aeration. 3. Mix cells for 15 h at 30°C on a platform shaker and then induce by addition of 0.4 mL of 3× YEP-GAL (see subheading 2.3, item 1) for 6 h, and follow by centrifugation and washing with ice cold water and storage at −80°C. 4. Obtain the lysates by adding 200 mL of lysis buffer 150 (see subheading 2.3, item 2) and shaking for 6 min in a paint shaker (5G-HD, Harbil) at 4°C with 250 mL of acid-washed

Protein Microarrays

217

glass beads (0.5 mm, Sigma), followed by centrifugation at 1,000 ´ g for 5 min. 5. Combine the crude lysates with 5× SDS loading buffer in a 96-well PCR plate, heat for 5 min at 95°C, and centrifuge again at 2,500 rpm 5 min. 6. Remove 12 mL from each well with a multichannel pipette and load onto SDS-PAGE gels. 7. Western transfer to either PVDF or nitrocellulose membranes using standard wet or dry transfer techniques. 8. Block the membrane with TBS/0.1% Tween–1% milk powder for 1 h. 9. Probe with anti-HA antibodies (16B12, 1:1,000, Covance) overnight in TBS/0.1% Tween–1% milk powder. 10. On day 2, wash the membrane three times with TBS/0.1% Tween, 10 min each wash. 11. Probe for 1 h at room temperature with HRP-conjugated sheep anti-mouse IgG antibody (Amersham). 12. Wash the membrane three times with TBS/0.1% Tween, 10 min each wash, and develop with Supersignal west chemiluminescent substrate (Pierce). Purifying Proteins for Arrays

1. Inoculate 96-well starter cultures in 0.8 mL of Sc-ura medium. 2. On day 2, wash the cells with Sc-ura/2% raffinose, and inoculate 5 mL of starter culture into a new 96-well box (2 mL/well) containing 0.8 mL of Sc-ura/2% raffinose with a 3.5-mm glass ball (PGC scientific) in each well for mixing and aeration. 3. Mix cells for 15 h at 30°C on a platform shaker and then induce by addition of 0.4 mL of 3× YEP-GAL (see Subheading 2.3, item 1) for 6 h followed by centrifugation, washing with ice cold water, and storage at −80°C. 4. Obtain the lysates by adding 200 mL of lysis buffer 150 (see Subheading 2.3, item 2) and shaking for 6 min in a paint shaker (5G-HD, Harbil) at 4°C with 250 mL of acid-washed glass beads (0.5 mm, Sigma), followed by centrifugation at 2,500 rpm for 5 min. 5. Transfer the cleared lysate to a new 96-well box along with IgG Sepharose resin (6 Fast Flow; GE Biosciences) and incubate at 4°C for several hours with agitation. 6. Afterwards, wash the beads four times with wash buffer 150 (see Subheading 2.4, item 1) 7. Transfer the beads to a PVDF filter plate (1.2 µm pore size) and spin-through the excess wash. Suspend the resin is in 40 µL of elution buffer containing1 µL of 3C protease for 18 h at 4°C with agitation.

218

Fasolo and Snyder

8. The next day, add preconditioned glutathione beads to the reaction to remove the 3C protease and spin the eluted protein through the PVDF membrane into a clean 96-well freezer plate. 9. Array the proteins into 5 mL aliquots on multiple 384-well printing plates and cover with an adhesive aluminum seal for storage at −80°C until printing. 10. Include a control plate with assay-specific controls (see Note 2). 3.1.4. Printing Protein Microarrays

Once proteins are prepared they must be printed onto surfaces, typically microscope slides. A variety of different surfaces exist for preparing protein microarrays. They include nitrocellulose and aldehyde surface chemistries for chemical attachment through lysines and affinity attachment methods such as nickelchelated slides for attaching His tagged proteins and glutathione for affinity attachment (15, 3, 16). The different surfaces have advantages and disadvantages; nitrocellulose allows the attachment of large amounts of protein, and it along with chemical attachment methods results in random orientation of proteins away from the surface. It is likely that the activity of the attached molecule is lost for many molecules attached by these methods. Moreover, binding of probes to regions near the surface is likely to be sterically inhibited. Affinity attachment methods are more likely to retain activity and orient proteins away from the surface, although random presentation of protein surfaces from the slide is not attained. Different surfaces yield different backgrounds and it is prudent to test several surfaces whenever a new assay is employed. 1. Configure the contact arrayer and align the microscope slides (FAST, Path, or other types of slides see Note 1) to allow the maximum number of slides to be printed per run. This will vary depending on the arrayer used, but for the Bio-Rad ChipWriter Pro it is approximately 90 arrays/run. 2. Maintain a constant humidity of ~40% in the printing chamber, and a constant temperature of 4°C (this is achieved by placing the printer in a temperature-controlled cold room). 3. Remove two plates at a time from the −80°C storage, and allow them to thaw on ice for 5–10 min followed by centrifugation at 1,000 × g at 4°C for 2 min. 4. Remove the foil seal from the top of the plate and place the plate into the chamber in the proper orientation for printing (see manufacturer’s instructions). 5. After the print run, cover the plate with a new aluminum seal and return to the −80°C freezer.

Protein Microarrays

219

6. Repeat for each plate. 7. After printing, store the arrays in a −20°C non-deicing freezer, which can remain stable for up to 1 year. 3.1.5. Probing for Protein– Protein Interactions with Protein Microarrays

3.1.6. Purifying the V5-Fusion Protein Probe

Protein microarrays can be screened for a wide variety of activities. A common activity is interaction with other proteins. One of the best ways to perform this is to produce the protein of interest with an epitope tag (we prefer to use the V5 epitope for yeast) and probe a yeast proteome chip as follows: 1. Grow 5–20 mL of yeast culture containing V5-fusion protein (pYES-DEST52 vector, Invitrogen, see Note 4) probe overnight in Sc-ura/2% dextrose. 2. Inoculate 40–400 mL of the Sc-ura/2% raffinose culture with sufficient starter culture to a final OD600 of 0.1. 3. Grow a large culture for a period of three doubling times and induce with 3×-YEP supplemented with 2% galactose, by adding enough to dilute the induction media by a factor of 3. 4. Induce cells at 30°C for 5 h. 5. Harvest cells using in a JA-10 (or comparable) rotor by spinning £400 mL volumes of cell suspension at 4,000 rpm/4°C/5 min. 6. Wash the cells once with 50 mL of cold dH2O and transfer to a 50-mL conical tube. Wash again in cold buffer (without detergents or other additives) used for lysis (e.g., PBS, Tris, HEPES) and transfer to 2-mL snap-cap tubes for lysis. 7. Spin the cells at 20,000 × g/4°C/1 min to a pellet and pipette away the buffer. 8. Place tubes on ice and proceed with lysis step. 9. Lyse the cells with 0.5-mm zirconia beads in a 1:1:1 volumes of cell pellet, beads, and PBS lysis buffer (see Subheading 2.6, item 1). Vortex the mixture using a paint shaker several times at 2-min intervals at 4°C. 10. Centrifuge the lysate at 20,000 × g in a tabletop microfuge for 10 min at 4°C, followed by ulracentrufugation for 30 min at 150,000g at 4°C. 11. Apply the clarified lysate to the previously mentioned affinity resin. 12. Wash the resin several times with wash buffer (see Subheading 2.6, item 5), followed by elution with elution buffer (see Subheading 2.6, item 6)

Probing the Arrays

1. Dilute the protein probe over a concentration range of 5–500 µg/mL, which must be optimized for each protein– protein interaction assay (see Fig. 1). The purified V5-fusion

220

Fasolo and Snyder

Print Arrays

Probe several arrays with anti-HA (or anti-GST) Antibody

Consistent Spotting Efficiency No

Probe arrays With purified V5-protein Probe (50 500 mg/mL)

Check pixel intensity Of Positive V5 control And Ab signals

Analyze data with Genepix or other Array analysis software

Titer Antibody Or Probe Concentration

Subtract out Negative control

Validate Interactions In vivo

Pixel intensity Within linear Range of detection No

Yes

Yes

Fig. 1. Flowchart for protein–protein interaction assay optimization.

protein is diluted into the probe buffer (see Subheading 2.7, item 2). 2. Remove the arrays from the freezer (−20°C) and bring to 4°C in the refrigerator, just prior to use, for about 15–20 min. Block the arrays in blocking buffer (see Subheading 2.7, item 1) for 1 h by shaking at 50 rpm on a stage at 4°C. 3. After blocking, transfer the arrays to a humidity chamber, and add 90 µL of diluted probe directly to the array surface. Overlay the arrays with a raised lifter slip and incubate static (no shaking) in the humidor for 1.5 h. 4. Wash the arrays 3 times for 1 min each in probe buffer in three 50-mL conical tubes (see Note 4). 5. To detect interactions, dilute the V5-AlexaFluor 647 antibody to 260 ng/mL in probe buffer and mix thoroughly by shaking. 6. After washing the arrays several times as indicated, add antibody solution directly to array and overlay with a raised lifter slip as before. Incubate the arrays for 30 min/static/4°C. 7. Finally, wash the arrays for 1 min (3×) in probe buffer, and spin in a 50-mL conical tube at 800g in a tabletop centrifuge for 5 min at room temperature. Air-dry the arrays in a slide holder in the dark for 30 min prior to scanning the array at 647 nm.

Protein Microarrays 3.1.7. Identifying Protein Kinase Substrate with Protein Arrays

221

Another useful application of protein microarrays is to identify targets of protein modification enzymes. Recently, a large study has been performed to identify the targets of 87 distinct protein kinases of yeast (12). A protocol for screening for targets of protein kinases is as follows: 1. Grow cells in 50–500-mL cultures, harvest, and lyse, with glass beads in the lysis buffer (see Subheading 2.7, item 1) as in “Purifying Proteins for Arrays.” 2. Kinases-GST fusions are bound to glutathione beads and eluted into the kinase buffer (see Subheading 2.7, item 2). 3. Block the proteome arrays in a Superblock (Pierce) with 0.1% Triton X-100 for 1 h at 4°C and probe in duplicate for every kinase. 4. Optimize conditions: Dilute the kinase into kinase buffer plus 0.5 mg/mL BSA, 0.1% Triton X-100, and 2 L of [33P]ATP (33.3 nM final concentration). 5. Overlay each kinase in buffer on two arrays, cover with a coverslip, and place in a humidified chamber at 30°C for 1 h. 6. Wash the slides twice with 10 mM Tris–HCl pH 7.4, 0.5% SDS and once with double-distilled water before being spun dry and exposed to X-ray film (Kodak). 7. For each experiment, incubate two additional arrays with kinase buffer in the absence of kinase, which will serve as autophosphorylation reference slides.

4. Notes 1. Care needs to be taken when printing samples with high glycerol content on surfaces other than nitrocellulose (nickel, ultrgap, etc.) in high humidity >40% because the spot diameter will differ and this could result in samples bleeding together. 2. Probe-specific control spots must be added to each array. These include known concentrations of 3C protease (if used) as well as Alexa Fluor V5 antibody and V5-fusion constructs. These will be used to ascertain optimal probing conditions and antibody titer for protein probing experiments. 3. Primers for sequencing of V5-fusions in pYES-DEST52 can be obtained from Invitrogen’s Website. ORFs can be shuttles from any of the destination (DEST) vectors using the pDONR221 as an intermediary. 4. When probing for protein–protein interactions, it is advisable to wash the arrays using three separate 50-mL conical tubes

222

Fasolo and Snyder

per slide. Allow the slides to remain submerged in the first chamber until the lifter slip falls off (do not pull it off with force). Afterward, proceed with 1 min washes, gently removing slides with forceps from each bath by gripping the nonmembraneous portion of the slide. References 1. Zhu H, et al. (2003) Proteomics. Annu. Rev. Biochem. 72:783–812. 2. Zhu H, Snyder M. (2003) Protein chip technology. Curr. Opin. Chem. Biol. 7(1): 55–63. 3. Zhu H, et al. (2001) Global analysis of protein activities using proteome chips. Science 293(5537):2101–5. 4. Popescu SC, et al. (2007) Differential binding of calmodulin-related proteins to their targets revealed through high-density Arabidopsis protein microarrays. Proc. Natl. Acad. Sci. U S A 104(11):4730–5. 5. Hudson ME, et al. (2007) Identification of differentially expressed proteins in ovarian cancer using high-density protein microarrays. Proc. Natl. Acad. Sci. U S A 104(44):17494–9. 6. Zhu H, et al. (2006) Severe acute respiratory syndrome diagnostics using a coronavirus protein microarray. Proc. Natl. Acad. Sci. U S A 103(11):4011–6. 7. Rolfs A, et al. (2008) Production and sequence validation of a complete full length ORF collection for the pathogenic bacterium Vibrio cholerae. Proc. Natl. Acad. Sci. U S A 105(11):4364–9. 8. Huang J, et al. (2004) Finding new components of the target of rapamycin (TOR) signaling network through chemical genetics and

9.

10.

11.

12.

13.

14.

15.

16.

proteome chips. Proc. Natl. Acad. Sci. U S A 101(47):16594–9. Hall DA, et al. (2004) Regulation of gene expression by a metabolic enzyme. Science 306(5695):482–4. Ho SW, et al. (2006) Linking DNA-binding proteins to their recognition sequences by using protein microarrays. Proc. Natl. Acad. Sci. U S A 103(26):9940–5. Michaud GA, et al. (2003) Analyzing antibody specificity with whole proteome microarrays. Nat. Biotechnol. 21(12):1509–12. Ptacek J, et al. (2005) Global analysis of protein phosphorylation in yeast. Nature 438(7068):679–84. Gupta R, et al. (2007) Ubiquitination screen using protein microarrays for comprehensive identification of Rsp5 substrates in yeast. Mol. Syst. Biol. 3:116. Gelperin DM, et al. (2005) Biochemical and genetic analysis of the yeast proteome with a movable ORF collection. Genes Dev. 19(23):2816–26. MacBeath G, Schreiber SL. (2000) Printing proteins as microarrays for highthroughput function determination. Science 289(5485):1760–3. Ramachandran N, et al. (2004) Self-assembling protein microarrays. Science 305(5680):86–90.

Chapter 13 Analysis of Protein–Protein Interactions Using Array-Based Yeast Two-Hybrid Screens Seesandra V. Rajagopala and Peter Uetz Summary The yeast two-hybrid system (Y2H) is a powerful tool to identify protein–protein interactions. Here we describe array-based two-hybrid methods that use defined libraries of open reading frames (ORFs) as opposed to random genomic or cDNA libraries. The array-based Y2H system is well suited for interactome studies of existing ORFeomes or subsets thereof, preferentially in a recombination-based cloning system. Array-based Y2H screens efficiently reduce false positives by using built-in controls, retesting, and evaluation of background activation. Hands-on time and the amount of used resources grow exponentially with the number of tested proteins; this is a disadvantage for large genome sizes. For large genomes, random library screen may be more efficient in terms of time and resources, but not as comprehensive as array screens, and they require an efficient sequencing facility. However, large array screens require some extent of automation although they can be carried out manually on smaller scales. Future-generation Y2H plasmid constructs including tightly regulated expression systems and features that facilitate biochemical characterization will provide more efficient and powerful tools to identify interacting proteins. Key words: Yeast two-hybrid system, Protein–protein interactions, Two-hybrid array

1. Introduction Specific interactions between proteins form the basis of many essential biological processes. Comprehensive analysis of protein– protein interactions is a challenging task of proteomics and has been best explored in budding yeast. Protein interactome analysis at a genome scale was first achieved by using yeast two-hybrid (Y2H) screens (1) and next by large-scale mass spectrometric analysis of affinity-purified protein complexes (2, 3). The Y2H

I. Stagljar (ed.), Yeast Functional Genomics and Proteomics, Methods in Molecular Biology, vol. 548 DOI: 10.1007/978-1-59745-540-4_13, © Humana Press, a part of Springer Science + Business Media, LLC 2009

223

224

Rajagopala and Uetz

system is a genetic method that detects binary protein–protein interactions in vivo. Classical two-hybrid screens used random libraries (genomic or cDNA) to identify novel interactions for a protein of interest. However, more recently, an array-based variation of this original principle has been increasingly used (Fig. 1). This approach can be applied to a few proteins but also to whole genomes. Advantages of arrays are their built-in controls and their systematic nature. However, random library screens may be more comprehensive and may yield fewer false negatives. Protocols for random library screens are not included in this chapter but may be found in (4). 1.1. The Principle of the Yeast Two-Hybrid System

The Y2H system is a genetic method extensively used to detect binary protein–protein interactions in vivo (in yeast cells). This system was developed by Stanley Fields (5) based on the observation that protein domains can be separated and recombined and still can retain their properties. In particular, transcription factors can frequently be split into the DNA-binding domain (DBD) and activation domains (ADs). In the two-hybrid system, a DNAbinding domain (here: from the yeast Gal4 protein) is fused to a protein “B” (for bait) for which one wants to find interacting partners (Fig. 1, step 3). A transcriptional activation domain is then fused to some or all the predicted open reading frames (ORFs or “preys”) of an organism. Bait and prey fusion proteins are then co-expressed in the same yeast cell. Usually both protein fusions are expressed from plasmids that can be manipulated easily and then transformed into yeast cells. If the two proteins “B” and ORF interact, a transcription factor is reconstituted, which in turn activates one or more appropriate reporter genes. The expression of the reporter allows the cell to grow only under certain conditions. For example, the HIS3 reporter encodes imidazoleglycerolphosphate (IGP) dehydratase, a critical enzyme in histidine biosynthesis. In the Y2H screening strain lacking an endogenous copy of HIS3, expression of a HIS3 reporter gene is driven by a promoter that contains a Gal4p-binding site, so the bait protein fusion can bind to it. However, since the bait fusion does not contain a transcriptional activation domain it remains inactive. If a protein ORF with an attached activation domain binds to the bait, this activation domain can recruit the basal transcription machinery, and expression of the reporter gene ensues. These cells can now grow in the absence of histidine in the media because they can synthesize their own.

1.2. Applications

Originally, the two-hybrid system was invented to demonstrate the association of two proteins (5). Later, it was demonstrated that completely new protein interactions can be identified with this system, even when there are no candidates for an interaction with a given bait. Over time, it has become clear that the ability to

Analysis of Protein–Protein Interactions

225

Fig. 1. Scheme of an array-based two-hybrid screen. Step 1: Yeast mating is the first step in the Y2H screening, which combines the bait and prey plasmids. First the bait (DNA-binding domain (DBD) fusion) liquid culture is pinned onto YEPD agar plates using a 384-pin pinning tool, and then the prey array (activation-domain (AD) fusion) is pinned on top of the baits using the sterile pinning tool. Then the mating plates are incubated at 30°C for 16 h. Step 2: The yeast mating plates are pinned onto –Leu –Trp (–LT) medium plates using a sterile 384-pin pinning tool. On –LT plates only diploid cells will grow. Selection on –LT media ensures that both the prey and bait plasmids are combined in the diploid yeast cells. Step 3: The diploid cells are pinned onto –Leu –Trp –His medium plates for protein interaction detection. Only if bait and prey proteins interact, and an active transcription factor is reconstituted transcription of a reporter gene is activated (lower right panel). In the array-based Y2H system, a bait protein is tested against the whole systematic prey library. This library consists of individual yeast colonies at specific positions of an array (e.g., in 384 format); each colony carries a specific prey construct of a specific ORF. Systematic testing is done by robotic transfer of yeast cells starting with the mating of bait and prey strains. Diploid cells are selected on specific plates and, finally, transferred to plates selecting for the activation of a reporter gene. Here the activation of the His3p gene is detected by scoring the growth on His-deficient plates. The rectangles on the selective plate mark negative and positive interactions between the bait and the prey at these specific positions of the array (test is done in duplicates). After Uetz et al. (1).

226

Rajagopala and Uetz

conveniently perform unbiased library screens is the most powerful application of the system. With whole-genome arrays, such unbiased screens can be expanded to defined, nonredundant sets of proteins. Arrays, like traditional two-hybrid screens, can also be adapted to a variety of related questions, such as the identification of mutants that prevent or allow interactions (6), the screening for drugs that affect protein interactions (7, 8), the identification of RNA-binding proteins (9), or the semiquantitative determination of binding affinities (10). The system can also be exploited to map binding domains (11), to study protein folding (12), or to map interactions within a protein complex (proteosome (13), flagellum (14)). Finally, recent large-scale projects have been successful in systematically mapping interactions within whole proteomes or subsets thereof (yeast (1, 15); worm (16); fly (17); human (18, 19)). These studies have shown for the first time that most proteins in a cell are actually connected to each other (20). In combination with structural genomics, gene expression data, and metabolic profiling, the enormous amount of data in these interaction networks should allow us eventually to model complex biological phenomena in molecular detail. An ultimate goal of this work is to understand the interplay of DNA, RNA, and proteins, together with small molecules, in a dynamic and realistic way. 1.3. Limitation of Yeast Two-Hybrid Screening

False negatives: Two-hybrid screens are not perfect. It is quite unlikely that you will detect all physiologically relevant interactors of your bait protein. False negatives may arise from steric hindrance of the two fusion proteins, so that physical interaction or subsequent transcriptional activation is prevented. Other explanations for false negatives include instability of proteins or failure of nuclear localization; the absence of a prey protein from a library; and inappropriate post-translational modification of a bait or prey, prohibiting an interaction. We estimate that the false negative rate in array-based two-hybrid screens is on the order of 75%; i.e., up to 75% of all “true” interactions may be missed (14). This large number can be reduced by several strategies. For example, we investigated the interactome of bacterial flagella by using ORFs from both Treponema pallidum and Campylobacter jejuni which had estimated false negative rates of 76% and 77%, respectively. However, a combination of both datasets recovered 33% of all known flagellar interactions and thus had a false negative rate of 67%. When protein domains and fragments are used, this number can be further reduced, although it may be difficult to recover more than 50% of all interactions using the Y2H. False positives: Like many assay systems, the two-hybrid system has the potential to produce false positives. “False positives” may be of technical or biological nature. The technical false positive is an apparent two-hybrid interaction that is not based on

Analysis of Protein–Protein Interactions

227

the assembly of two hybrid proteins (that is, the reporter gene(s) gets activated without a protein–protein interaction between bait and prey being involved). Frequently, such false positives are associated with bait proteins that act as transcriptional activators. Some bait or prey proteins may affect general colony viability and hence enhance the ability of a cell to grow under selective conditions and activate the reporter gene. Mutations or other random events of unknown nature may be invoked as potential explanations as well. A number of procedures have been developed to identify or avoid false positives, including the utilization of multiple reporters, independent methods of specificity testing, or simply repeating assays to make sure a result is reproducible (21–23) (described below). The biological false positive means a bona fide two-hybrid interaction with no physiological relevance. Those include the partners that can physically interact but that are never in close proximity to one another in the cell because of distinct subcellular localization or expression at different times during the life cycle of a cell or organism. Examples may include paralogs that are expressed in different tissues or at different developmental stages. The problem is that the “false positive” nature can rarely be proven, as there may be unknown conditions under which these proteins do interact with a biological purpose. Overall, hardly any false positives can be explained mechanistically (although many may simply do interact nonphysiologically!). While it often remains difficult to prove the biological significance of an interaction, many studies have attempted to validate them by independent methods. Finding an interaction by several methods certainly increases the probability that it is biologically significant. Recently, Uetz et al. (24) evaluated the all the Y2H interactions of Kaposi Sarcoma Herpes virus (KSHV) by co-immunoprecipitation (coIP) and found that about 50% of them can be confirmed. Similarly, when subsets of the large-scale human Y2H interactome were evaluated, 78 and 65% of them could be verified by independent methods (18, 19). 1.4. Array-Based Screens

In an array, a number of defined prey proteins are tested for interactions with a bait protein (Fig. 1). Usually, the bait protein is expressed in one yeast strain and the prey is expressed in another yeast strain of different mating type. The two strains are then mated so that the two proteins are expressed in the resulting diploid cell (Fig. 1). The assays are done side-by-side under identical conditions, so they can be well controlled, i.e., compared. As the identity of the preys is usually known, no sequencing is required after positives have been identified. However, the prey clones need to be obtained or made upfront. This can be done for a few genes or for a whole genome, e.g., an ORFeome (i.e., all ORFs of a genome).

228

Rajagopala and Uetz

In an array, each element has a known identity, and therefore it is immediately clear which two proteins are interacting when positives are selected. In addition, it is often immediately clear if an interaction is stronger than another one (but see below). Most importantly, since all these assays are done in an ordered array, background signals can be easily distinguished from true signals (Fig. 1, step 3). Until recently, it was much easier to construct a random library and screen it rather than to construct many individual clones and screen them individually. However, now whole genomes become increasingly available as ordered clone sets in a variety of vectors. Modern cloning systems also allow direct transfer of entry clones into many specialized vectors (see below). For most model organisms such genome-scale clone collections are already available or will be soon. One of the first applications of such clone collections is often a protein interaction screen, so it is likely that a prey library is already available for your favorite organism! In fact, in some cases only an array screen may do the job. For example, if you have a bait protein that activates transcription on its own, a carefully controlled array may be the only way to distinguish between signal and background (see Fig.1, step 3). Similarly, weak interactors may be detectable only when compared with a uniformly weak or no background. 1.5. Pooled Array Screening

A completely different screening strategy, the pooling strategy, has the potential to accelerate screening a lot but might also have the disadvantage of increasing the number of false negatives. This may have been a reason why pooled screens in Campylobacter jejuni resulted in more false negatives than in the one-by-one screens of Treponema pallidum(14, 25). In the first step, sets of proteins (pools, rather than single proteins) are tested for interactions against each other (for example, pools of 10 known preys in each colony of the prey array). In the second step, positive sets are taken and the proteins defining this set are individually tested for their binary interactions (as in the classical array-based Y2H strategy). Depending on the pool size, the first level of screening is therefore usually fast and, as only a few interactions are expected for each protein, only a few pools need to be tested for binary interactions in the second step. Such a pooling strategy was established by Zhong et al.(26). Pools of prey proteins were tested against single bait proteins, and it was shown that the pools could contain 96 or even more proteins. The authors calculated that the Y2H array screening of the whole yeast genome (~6,000 proteins) would require only 1/24 of the time and effort when using the proposed pooling strategy as opposed to a one-by-one strategy. Prey-based pools are advantageous over bait-based pools because 10–20% of bait– BD fusions can activate the two-hybrid reporter gene (for example,

Analysis of Protein–Protein Interactions

229

the HIS3 gene) without the presence of any prey–AD fusion. Addition of 3-amino-triazole, an enzymatic inhibitor of His3, can inhibit autoactivation. Thus, it is possible to optimize twohybrid selection conditions based on the autoactivation level of the individual bait (by varying 3-amino-triazole concentration in the medium). Recently Jin et al. developed a smart pool-array (SPA) system (27) where they show that the pooling-deconvolution principle can be applied to pool prey (AD) strains, permitting efficient screening of individual baits with high accuracy and coverage. In an SPA scheme, 16 (=24) strains are mixed into four pairs of pools (pairs 0–3), with eight (=23) strains per pool. This pooling scheme has two important properties of deconvolution and redundancy. First, deconvolution is possible because every strain is pooled into four different pools (one from each pair), so if one of the 16 strains is two-hybrid positive (for a given bait), then four of the eight pools will yield a positive colony. Thus we can deconvolute the identity of the two-hybrid positive strain owing to its presence only in a specific combination of four pools and absence in the other pools. The second important property of SPA is the built-in redundancy; each AD strain is tested four times against the bait, resulting in a situation that is equivalent to four separate individual screens. This inherent replication can facilitate removal of false positives because false positives are unlikely to be observed reproducibly. Likewise, replicated screens will cover more true positives, because losing the same true positive repeatedly as a result of experimental variation is also less likely. In general, this scheme generates 2n pools from 2n strains, with a built-in “screen redundancy” of n. Thus, when the number of strains is increased exponentially, for example, from 16 (=24) to 32 (=25), the number of pools only needs to be increased linearly, that is, from 8 (= 2 × 4) to 10 (= 2 × 5). 1.6. General Requirements for a Screen and Alternatives

Although the protocols in this chapter are based on the DNAbinding and activation domain of the yeast Gal4 protein, other DNA-binding domains and activation domains can be used. In the LexA two-hybrid system, the DNA-binding domain is provided by the entire prokaryotic LexA protein, which normally functions as a repressor in E. coli when it binds to LexA operators. In the Y2H system, the LexA protein does not act as a repressor, as the promoter with its binding sites is not constitutively active. An activation domain often used in the LexA two-hybrid system is the heterologous 88-residue acidic peptide B42 that strongly activates transcription in yeast. An interaction between the target protein (fused to the DNA-BD) and a library-encoded protein (fused to an AD) creates a novel transcriptional activator with binding affinity for LexA operators.

230

Rajagopala and Uetz

In general, every component of the “classic” two-hybrid system can be replaced by different components: For example, the reporter gene does not need to be HIS3. Alternatively, LEU2, an enzyme involved in leucine biosynthesis, can be used. The reporter does not have to be a biosynthetic enzyme at all; green fluorescent protein (GFP) has been successfully used as a reporter gene (28), beta-galactosidase (lacZ) is common (29), and many others are under investigation. Finally, the two-hybrid system does not need to be based on transcription. Johnsson and Varshavsky (1994) developed a related system that is based on reconstituting artificially split ubiquitin, a protein that tags other proteins for degradation. As long as the function of a protein can be used as a selective marker, it is theoretically possible to divide it into fragments, and drive the reassociation of the two fragments by exogenous “Bait and Prey” proteins which are attached to each half. Several other variations have been developed and are described elsewhere (30, 31). 1.7. Genome-Wide Yeast Two-Hybrid Screening (See Table 1)

The construction of an entire proteome array of an organism that can be screened in vivo under uniform conditions is a challenge. When proteins are screened at a genome scale, automated robotic procedures are necessary (see below). The procedure can be modified for manual use or for use with alternative screenings strategies such as synthetic lethal screens. With minor modifications, the array can be used to screen for protein interactions with DNA, RNA, or even small-molecule inhibitors of the proteinprotein interactions. The protocols described here were established for yeast proteins, but they can be applied to any other genome or subset thereof; for example, viral and bacterial genomes have been screened for interactions in our lab. Different high-throughput

Table 1 What you need for a yeast two-hybrid screen (examples) Bait plasmid(s)

pGBKT7 (Clontech), pOBD2, pDEST22 (Invitrogen), pGBKT7g (24)

Prey plasmid(s)

pGADT7 (Clontech), PAS1, pDEST22 (Invitrogen), pGADT7g (24)

Bait yeast strain

AH109

Prey yeast strain

Y187

Yeast media

YEPD, selective liquid media and agar plates

Pin tool

Optional but necessary when large numbers are tested

Analysis of Protein–Protein Interactions

231

cloning methods used to generate two-hybrid clones, i.e., proteins with AD fusions (preys) and the DBD fusions (baits) are therefore included below. The process involves the construction of the prey and bait array (described in Subheadings 3.2–3.4) and screening of the array by either manual or robotic manipulation (described in Subheading 3.5.1–3.5.4), including the selection of positives and scoring of results. High-throughput screening projects deal with a large number of proteins; therefore hands-on time and amount of resources become an important issue. Options to reduce the screening effort are discussed. A prerequisite for array-based genome-wide screen is the existence of a cloned ORFeome; we will briefly mention strategies to create such ORFeomes. Many ORFeome projects are currently being done. We expect readily available complete ORFeomes for all major model organisms in the near future.

2. Materials 2.1. Yeast Media

1. YEPD liquid medium: 10 g yeast extract, 20 g peptone, 20 g glucose. Make up to 1 L with sterile water, and autoclave. 2. YEPD solid medium: 10 g yeast extract, 20 g peptone, 20 g glucose, 16 g agar. Make up to 1 L with sterile water and autoclave. After autoclaving, cool the medium to 60–70°C, then add 4 ml of 1% adenine solution (1% in 0.1 M NaOH), pour 40 ml into each sterile Omnitray plate (Nunc) under sterile hood, and let them solidify. 3. Medium concentrate: 8.5 g yeast nitrogen base, 25 g ammonium sulfate, 100 g glucose, 7 g dropout mix (see below). Make up to 1 L with sterile water, and filter-sterilize (Millipore).

2.2. Yeast Minimal Media (Selective) Plates

1. For 1 L of selective medium, autoclave 16 g agar in 800 ml water, cool the medium to 60–70° C, and then add 200 ml medium concentrate. Depending on the required selective plates, you have to add the missing amino acids or 3AT (3-amino1,2,4-triazole). 2. –Trp plates (media lacking tryptophan): Add 8.3 ml leucine and 8.3 ml histidine from the stock solution (see below). 3. –Leu plates (media lacking leucine): Add 8.3 ml tryptophan and 8.3 ml histidine solution from the stock solution. 4. –Leu -Trp plates (media lacking tryptophan and leucine): Add 8.3 ml histidine from the stock solution.

232

Rajagopala and Uetz

5. –Leu –Trp –His plates (media lacking tryptophan, leucine, and histidine): Nothing needs to be added. 6. –Leu –Trp -His + 3 mM 3AT plates: Add 6 ml of 3AT (3-amino-1,2,4-triazole, 0.5 M) to a final concentration of 3 mM. 7. Dropout mix (-His, -Leu, -Trp): Mix 1 g methionine, 1 g arginine, 2.5 g phenylalanine, 3 g lysine, 3 g tyrosine, 4 g isoleucine, 5 g glutamic acid, 5 g aspartic acid, 7.5 g valine, 10 g threonine, 20 g serine, 1 g adenine, and 1 g uracil and store under dry, sterile conditions. 8. Amino acid stock solutions: Histidine (His): dissolve 4 g of histidine in 1 L sterile water and filter-sterilize. Leucine (Leu): dissolve 7.2 g of leucine in 1 L sterile water and sterile filter. Tryptophan (Trp): dissolve 4.8 g of tryptophan in 1 L sterile water and filter-sterilize. 2.3. Yeast Transformation

1. Salmon sperm DNA (Carrier DNA): Dissolve 7.75 mg/ml salmon sperm DNA (Sigma) in sterile water, autoclave for 15 min at 121°C, and store at −20°C. 2. Dimethylsufoxide (DMSO, Sigma). 3. Competent host yeast strains, e.g., AH109 (for baits), and Y187 (for preys). 4. Lithium acetate (LiOAc) (0.1 M). 5. Selective plates (depending on the selective markers, described in Subheading 2.2). 6. 96PEG solution: Mix 45.6 g PEG (Sigma), 6.1 ml of 2 M LiOAc, 1.14 ml of 1 M Tris-HCl, pH 7.5, and 232 ml 0.5 M EDTA. Make up to 100 ml with sterile water and autoclave. 7. Plasmid clones or linearized vector DNA and PCR product (for homologous recombination).

2.4. Bait Self Activation Test

1. YEPD liquid medium and selective media agar in single-well microtiter plates (Omnitray plates, Nunc). 2. –Trp –Leu plates (see Subheading 2.2). 3. Selective plates without Trp, Leu, and His, but with different concentrations of 3-AT, e.g., 0 mM, 1 mM, 3 mM, 10 mM, 50 mM and 100 mM (–LTH/3-AT plates). 4. Bait strains and the prey strain carrying the empty prey plasmid, e.g., Y187 strain with pDEST22 plasmid (Invitrogen).

2.5. Two-Hybrid Screening Protocol

1. 20% (v/v) bleach (1% sodium hypochlorite). 2. 95% (v/v) ethanol. 3. Single-well microtiter plates (e.g., OmniTray; Nalge Nunc) containing solid YEPD + adenine medium (see Subheading 2.1),

Analysis of Protein–Protein Interactions

233

–Leu –Trp (–LT), –His –Leu –Trp (–LTH), and –His –Leu –Trp + different concentrations of 3AT. 4. 384-Pin replicator for manual screening or robot (Biomek FX). 5. Bait liquid culture (DBD fusion-expression yeast strain). 6. Yeast prey array on solid YEPD plates. 2.6. Retest of Protein Interactions

1. 96-well microtiter plates (U- or V-shaped). 2. YEPD medium and YEPD agar in Omnitrays (Nunc). 3. Selective agar plates (–LT, –LTH with 3-AT). 4. Prey yeast strain carrying empty prey plasmid, e.g., pDEST22 in Y187 strain. 5. Bait and prey strains to be retested.

2.7. Beta-Galactosidase Filter Lift Assay

1. Selective plate (–LT) with diploid yeast colonies (from Subheading 3.5.4). The diploid cells carry the bait and prey combinations to be tested for activation of the beta-galactosidase reporter. 2. Omnitray plate. 3. Nitrocellulose membrane and Whatman paper. 4. Z-buffer: 60 mM Na2HPO4 (anhyd.), 60 mM NaH2PO4, 10 mM KCl, 1 mM MgSO4. 5. X-GAL solution: 40 mg/ml X-Gal (5-bromo-4-chloro-3indolyl-b-D-galactopyranoside) in dimethylformamide (DMF).

3. Methods 3.1. Strategic Planning

Before starting an array-based screen, the size and character of the array must be designed and the ultimate aims of the experiment need to be considered. Factors that may be varied include the form of protein array (e.g., full-length protein or single domain, choice of epitope tags etc.). Similarly, the arrayed proteins may be related (e.g., a family or pathway of related proteins, orthologs of a protein from different species, the entire protein compliment of a model organism). In our experience, certain protein families work extremely well (e.g., splicing proteins), while others do not appear to work at all (e.g., many metabolic enzymes). We recommend to carry out a small-scale pilot study, incorporating positive and negative controls, before committing to a full-scale project. Although high-throughput screening projects can be performed manually, automation is strongly recommended. Highly repetitive tasks are not only boring and straining but also errorprone when done manually. If you do not have local access to robotics, you may have to collaborate with a laboratory that has.

234

Rajagopala and Uetz

3.2. Generation of a Protein Array Suitable for High-Throughput Screening

Once the set of proteins to be included in the array is defined, the coding genes need to be PCR-amplified and cloned into Y2H bait and prey vectors. In order to facilitate the cloning of a large number to proteins, site-specific recombination-based systems are commonly used (e.g., Gateway (32), see Fig. 2b). Gateway cloning requires expensive enzymes and vectors, although both may be produced in the lab.

3.2.1. Cloning by Homologous Recombination in Yeast

An alternative to site-directed systems is the cloning by homologous recombination directly in yeast (33). A two-step PCR protocol is used to make DNA with sufficient homology to vector DNA at the terminal ends to allow homologous recombination in the yeast cell (Fig. 2a). In the “first round,” PCR reaction the ORF is amplified with primers that contain ~20-nucleotide tails which are homologous to sequences in the two-hybrid vectors. In the second-round PCR, ~50-nucleotide tails are attached to the first-round PCR products that are homologous to the destination vector cloning site (Fig. 2). The PCR product is then transformed into the yeast cells together with the linearized vector, and the recombination event between them takes place inside the yeast cell. The advantage of this strategy is its much reduced cost.

Fig. 2. Cloning strategies for creating baits and preys. (a) Homologous recombination. ORFs are amplified (first PCR) with specific primers that generate a product with common 5’ and 3’ 20-nucleotide tails. A second PCR generates a product with common 5’ and 3’ 70-nucleotide tails. The common 70-nucleotide ends allow cloning into linearized twohybrid expression vectors by co-transformation into yeast. The endogenous yeast recombination machinery performs the recombination reaction and results in a circular plasmid. (b) The Gateway-based Y2H expression clones are made by combining the ORFs of interest from a Gateway entry vector (such as pDONR201 or pDONR207) and the Y2H expression vectors (such as pDEST22 and pDEST32) in Gateway LR Clonase (Invitrogen) enzyme mix, which transfers the ORF from an entry clone to Y2H expression vectors.

Analysis of Protein–Protein Interactions

235

The disadvantage is that the resulting plasmids are not available as purified DNA but have to be recovered from yeast, which can be time consuming and inefficient. 3.2.2. Gateway Cloning

Gateway (Invitrogen) cloning provides another fast and efficient way of cloning the ORFs (32). It is based on the site-specific recombination properties of bacteriophage lambda (34); recombination is mediated between so-called attachment sites (att) of DNA molecules: between attB and attP sites or between attL and attR sites. The first step to Gateway cloning is inserting the gene of interest into a specific entry vector. One way of obtaining the initial entry clones is by recombining a PCR product of the ORF flanked by attB sites with the attP sites of a pDONR vector (Invitrogen). The resulting entry clone plasmid contains the gene of interest flanked by attL recombination sites. These attL sites can be recombined with attR sites on a destination vector, resulting in a plasmid for functional protein expression in a specific host (for example, pDEST22 and pDEST32, Fig. 2b).

3.2.3. The ORFeome

The starting point of an systematic array-based Y2H screening is the construction of an ORFeome. An ORFeome represents all ORFs of a genome – in our case, the selected gene set individually cloned into entry vectors. More and more ORFeomes are available and can be directly used for generating the Y2H bait and prey constructs. These ORFeome range from small viral genomes, e.g., KSHV and VZV (24), to several bacterial genomes such as Bacillus anthracis or Yersinia pestis. These bacterial ORFeomes are available from the Pathogen Functional Genomics Resource Center (http://pfgrc.jcvi.org/). Clone sets of multicellular eukaryotes, e.g., C. elegans(35), human (36), or plant (37), have also been described. However, not all genes of interest are already available in entry vectors. Both entry vector construction and the subsequent destination vector cloning can be done for multiple ORFs in parallel. The whole procedure can be automated using 96-well plates so that whole ORFeomes can be processed in parallel.

3.3. The Prey Array

The Y2H array is set up from an ordered set of AD-containing strains (preys) rather than BD-containing strains (baits), because the former do not generally result in self-activation of transcription. The prey constructs are assembled by transfer of the ORFs from entry vectors into specific prey vectors by recombination. Several prey vectors for the Gateway system are available. In our lab we use the Gateway-compatible pGADT7g vector, a derivative of pGADT7 (Clontech), or pDEST22 (Invitrogen) (Fig. 2b). We have also used pAS1 and other vectors for homologous recombination (Fig. 2a). These prey constructs are transformed into haploid yeast cells (described in Subheading 3.4); we use

236

Rajagopala and Uetz

the Y187 strain (mating type alpha) (Table 2). Finally, individual yeast colonies, each carrying one specific prey construct, are arrayed on agar plates in a 96- or 384-format in duplicates or quadruplicates. 3.3.1. Bait Construction

Baits are also constructed by recombination-based transfer of the ORFs into specific bait vectors or, alternatively, directly by homologous recombination in yeast. Bait vectors used in our lab are the Gateway-adapted pGBKT7g (Clontech) or pDEST32 (Invitrogen) for Gateway cloning and pOBD2 vector for homologous recombination cloning (Fig. 2a). The bait constructs are also transformed into haploid yeast cells (described in Subheading 3.4); we use the AH109 strain (mating type a) (Table 2). After self-activation testing, the baits can be tested for interactions against the Y2H prey array (see Note 1).

3.4. High-Throughput Yeast Transformation

This method is recommended for the high-throughput transformation of the bait or prey plasmid clones into respective yeast strains, and is optimized on the method of Cagney et al. (33). This protocol is suitable for 1,000 transformation, it can be scaled up and down as required, and most of the steps can be automated. Selection of the transformed yeast cells requires leucine- or tryptophan-free media (–Leu or –Trp depending on the selective marker on the plasmid). Moreover, at least one of the haploid strains must contain a two-hybrid reporter gene (here: HIS3 under GAL4 control). 1. Prepare competent yeast cells: Inoculate 250 ml YEPD liquid medium with yeast strains freshly grown on YEPD agar medium in a 2-L flask and grow in a shaker (shaking at 200 rpm) at 30°C. Remove the yeast culture from the shaker when the cell density reaches OD 1.0–1.3. This usually takes 12–16 h.

Table 2 Yeast strains and their genotypes Yeast strains

Genotypes

Y187

MATa, ura3- 52, his3- 200, ade2- 101, trp1- 901, leu2- 3, 112, gal4D, met–, gal80D,URA3::GAL1UAS -GAL1TATA -lacZ (after Harper et al. (39))

AH109

MATa, trp1-901, leu2-3, 112, ura352, his3-200, gal4D, gal80D, LYS2::GAL1UAS-GAL1TATA-HIS3, GAL2UAS-GAL2TATA-ADE2, URA3::MEL1UAS-MEL1 TATA-lacZ (after James et al. (40))

Analysis of Protein–Protein Interactions

237

2. Spin out the cells at 2,000 × g for 5 min at room temperature; pour off the supernatant. 3. Dissolve the cell pellet in 30 ml of LiOAc (0.1 M); make sure pellet is completely dissolved and there are no cell clumps. 4. Spin the cells in a 50-ml Falcon tube at 2,000 × g for 5 min at room temperature, pour off the supernatant, and dissolve the cell pellet in a total volume of 10 ml LiOAc (0.1 M). 5. Prepare the yeast transformation mix without yeast cells by mixing the components listed below in a 200-ml sterile bottle. Component

For 1,000 reactions

96PEG

100 ml

Salmon sperm DNA

3.2 ml

DMSO

3.4 ml

6. Add the competent yeast cells prepared above (steps 1–4) to the yeast transformation mix; shake the bottle vigorously by hand, or vortex for 1 min. 7. Pipette 100 ml of the yeast transformation mix into a 96-well transformation plate (we generally use Costar 3596 plates) by using a robotic liquid handler (e.g., Biomek FX) or a multistep pipette. 8. Now add 25–50 ng of plasmid; keep one negative control (i.e., only yeast transformation mix). 9. Seal the 96-well plates with plastic or aluminum tape and vortex for 2–3 min. Care should be taken to seal the plates properly; vigorous vortexing might cause cross-contamination. 10. Incubate the plates at 42°C for 30 min. 11. Spin the 96-well plate for 5 min at 2,000 × g; aspirate the supernatant and dry by tapping on a cotton napkin a couple of times. 12. Add 150 ml of selective liquid media to each well (depending on the selective marker on the plasmid construct (for example, trytophan- or leucine-free liquid media). Seal the plates with AirPore tape (Qiagen) to protect from evaporation. 13. Incubate at 30°C for 36–48 h. 14. Pellet the cells by spinning at 2,000 × g for 5 min, discard the supernatant, and add 10 ml sterile H2O to each well. 15. Transfer the cells to selective agar plate to select yeast with transformed plasmid (single-well Omnitrays 128 × 86 mm from Nunc are well suited for robotic automation). Typically,

238

Rajagopala and Uetz

we use a 96-pin tool (see reagent setup for the sterilization of the pin tool). As an alternative to the pin tool, one can use a multichannel pipette to transfer the cells. Allow the yeast spots to dry on the plates. 16. Incubate at 30°C for 2 days. Colonies start appearing after 24 h. 3.5. Screening and Retesting 3.5.1. Self-Activation Test

Prior to the two-hybrid analyses, the bait yeast strains should be examined for self-activation. Self-activation is defined as detectable bait-dependent reporter gene activation in the absence of any prey interaction partner. Weak to intermediate strength selfactivator baits can be used in two-hybrid array screens because the corresponding bait–prey interactions confer stronger signals than the self-activation background. In case of the HIS3 reporter gene, the self-activation background can be titrated by adding different concentrations of 3-AT, a competitive inhibitor of HIS3. Self-activation of all the baits is examined on plates containing different concentrations of 3-AT. The lowest concentration of 3-AT that suppresses growth in this test is used for the interaction screen (see below), because it avoids background growth while still detecting true interactions. The aim of this test is to measure the background reporter activity (here: HIS3) of bait proteins in absence of an interacting prey protein. This measurement is used for choosing the selection conditions used for Y2H screening described in Subheading 3.5.2. 1. Bait strains are arrayed onto a single-well Omnitray agar plate; either the standard 96-spot format or the 384-spot format is used (see Note 2). 2. The arrayed bait strains are mated with a prey strain carrying the empty prey plasmid, e.g., Y187 strain with pDEST22 (Invitrogen). Mating is conducted according to the standard screening protocol as described in Subheading 3.5.2. Note that here an array of baits is tested whereas in a “real” screen (Subheading 3.5.2) an array of preys is tested. 3. After selecting for diploid yeast cells (on –LT agar), the cells are transferred to media selecting for the HIS3p reporter gene activity as described in Subheading 3.5.2. The -LTH transfer may be done to multiple plates with increasing concentrations of 3-AT. Suggested 3-AT concentrations are 0, 1, 3, 10, 25, 50, and 100 mM. 4. These –LTH + 3-AT plates are incubated for 1 week at 30°C. The self-activation level of each bait is assessed: the lowest 3-AT concentration that completely prevents colony growth is noted. As this concentration of 3-AT

Analysis of Protein–Protein Interactions

239

suppresses reporter activation in the absence of an interacting prey, this 3-AT concentration is added to –LTH plates in the actual interaction screens as described in Subheading 3.5.2. 3.5.2. Screening for Protein Interactions Using a Yeast Protein Array

The Y2H prey array can be screened for protein interactions by a mating procedure that can be carried out manually or using robotics. A yeast strain expressing a single candidate protein as a DBD fusion is mated to all the colonies in the prey array (Fig. 1, step 1: shown for one prey plate). After mating, the colonies are transferred to a diploid-specific medium, and then to the two-hybrid interaction selective medium. To manually screen with more than one bait, replicate copies of the array are used. For large numbers of baits, robotic screening is recommended. In many cases, a hand-held 384-pin replicating tool can be used for routine transfer of colonies for screening. For large projects, however, a robotic workstation (e.g., Biomek 2000, or Biomek FX, Beckman Coulter) may be used to speed up the screening procedures and to maximize reproducibility. A 384-pin steel replicating tool (e.g., High-Density Replication Tool; V&P Scientific) can be used to transfer the colonies from one plate to another. Between the transfer steps, the pinning tool must be sterilized, see below). Note that not all plasticware is compatible with robotic devices, although most modern robots can be reprogrammed to accept different consumables. In the procedure described here, the prey array is gridded on 86 × 128 mm single-well microtiter plates (e.g., OmniTray, Nalge Nunc International) in a 384-colony format (see Fig. 1). 1. Sterilization: Sterilize a 384-pin replicator by dipping the pins into 20% bleach for 20 s, sterile water for 1 s, 95% ethanol for 20 s, and sterile water again for 1 s. Repeat this sterilization after each transfer. Note: Immersion of the pins into these solutions must be sufficient to ensure complete sterilization. When automatic pinning devices are used, the solutions need to be checked and refilled occasionally (especially ethanol which evaporates faster than the others). Day 1: 2. Preparing prey array for screening: Use the sterile replicator to transfer the yeast prey array from selective plates to singlewell plates containing solid YEPD medium and grow the array overnight in a 30°C incubator (see Note 3). 3. Preparing bait liquid culture (DBD fusion-expressing yeast strain): Inoculate 20 ml of liquid YEPD medium in a 250-ml conical flask with a bait strain and grow overnight in a 30°C shaker (see Note 4).

240

Rajagopala and Uetz

Day 2: 4. Mating procedure: Pour the overnight liquid bait culture into a sterile Omnitray plate. Dip the sterilized pins of the pin replicator (thick pins of ~1.5 mm diameter should be used to pin baits) into the bait liquid culture and place directly onto a fresh single-well plate containing YEPD agar media. Repeat with the required number of plates and allow the yeast spots to dry onto the plates for 10–20 min. 5. Pick up the fresh prey array (i.e., AD) yeast colonies with sterilized pins (thin pins of ~1 mm diameter should be used to pin the preys) and transfer them directly onto the baits on the YEPD plate, so that each of the 384 bait spots per plate receives different prey yeast cells (i.e., a different AD fusion protein). Incubate overnight at 30°C to allow mating (Fig.1 step 1, see Note 5). 6. Seletion of Diploids: For the selection of diploids, transfer the colonies from YEPD mating plates to plates containing –Leu –Trp medium using the sterilized pinning tool (thin pins should be used in this step). Grow for 2–3 days at 30°C until the colonies are >1 mm in diameter (Fig. 1 step 2, see Note 6). 7. Interaction selection: Transfer the colonies from –Leu –Trp plates to a single-well microtiter plate containing solid –His –Leu –Trp agar, using the sterilized pinning tool. If the baits are self-activating, they have to be transferred to –His –Leu –Trp + a specific concentration of 3AT (Subheading 3.5.1). Incubate at 30°C for 6–8 days. 8. Score the interactions by looking for growing colonies that are significantly above background by size and are present as duplicate colonies. 9. The plates should be examined every day. Most two-hybridpositive colonies appear within 3–5 days, but occasionally positive interactions can be observed later. Very small colonies are usually designated as background; however, there is no absolute measure to distinguish between the background and real positives. When there are many (e.g., >20) large colonies per array of 1,000 positions, we consider these baits as “random” activators. In this case, the screening should be repeated, or the interactions should be retested. 10. Scoring can be done manually or using automated image analysis procedures. When using image analysis, care must be taken not to score contaminated colonies as positives. 3.5.3. Protein Interaction Retesting

A major consideration when using the Y2H system is the number of false positives. The major sources for false positives are nonreproducible signals that arise through little-understood mechanisms. In Y2H screens, more than 90% of all interactions can be nonreproducible background (38). Thus, simple retesting by repeated

Analysis of Protein–Protein Interactions

241

mating can identify most false positives. We routinely use at least duplicate tests, although quadruplicates should be used if possible (see Fig. 1). Retesting is done by manually mating the interaction pair to be tested and by comparing the activation strength of this pair with the activation strength of a control, usually the bait mated with the strain that contains the empty prey vector. Testing for reproducibility of interactions greatly increases the reliability of the Y2H interaction data. This method is used for specifically retesting interaction pairs detected in an array screen. 1. Re-array bait and prey strains of each interaction pair to be tested into 96-well microtiter plates. Use separate 96-well plates for baits and preys. For each retested interaction, fill one well of the bait plate and one corresponding well of the prey plate with 150 ml YEPD liquid medium. 2. For each retested interaction, inoculate the bait strain into a well of the 96-well bait. 3. Plate bait and prey strain at the corresponding position of the 96-well prey plate, for example, bait at position B2 of the bait plate and prey at position B2 of the prey plate. In addition, inoculate the prey strain with the empty prey vector (e.g., strain Y187 with plasmid pDEST22) into 20 ml YEPD liquid medium. 4. Incubate the plates overnight at 30°C. 5. Mate the baits grown in the bait plate with their corresponding preys in the prey plate. In addition, mate each bait with the prey strain carrying an empty prey vector as a background activation control. The mating is done according to Subheading 3.5.2, using the bait and prey 96-well plates directly as the source plates (see Note 7). 6. The transfers to selective plates and incubations are done as described in Subheading 3.5.2. As before, test different baits with different activation strengths on a single plate and pin the diploid cells onto –LTH plates with different concentrations of 3-AT. For choosing the 3-AT range, the activation strengths (Subheading 3.5.1) serve as a guideline. 7. After incubating for ~1 week at 30°C on –LTH/3-AT plates, the interactions are scored; positive interactions show a clear colony growth at a certain level of 3-AT, whereas no growth should be seen in the control (bait mated with empty vector strain). 3.5.4. Beta-Galactosidase Filter Lift Assay (Alternative Reporter Genes)

Y2H interactions can be reproduced using other reporter genes in addition to the one used in the actual screen depending on the different reporter genes present in the yeast strains used. Examples include beta-galactosidase or ADE2 (for selection on adenine-deficient medium). Because of the use of different promoters, these reporter genes have different activation requirements, and Y2H interactions reproduced with different reporter

242

Rajagopala and Uetz

genes are assumed to be more reliable. However, the use of multiple reporters may result in the loss of weaker Y2H positives. The beta-galactosidase reporter has the advantage of giving a semiquantitative output of the activation strength. Other reporters might be advantageous and can be transformed into yeast as additional plasmids, or by using alternative strains which contain the reporter as integrated construct. For example, the strain AH109 carries an alpha-galactosidase reporter gene which produces an enzyme that is secreted into the medium. Therefore, these cells do not require cell lysis for detection. The following method was adapted from the Breeden lab (http://labs.fhcrc. org/breeden/Methods/index.html). 1. Use the same diploid plate as in Subheading 3.4.2. As a control, the bait strains are mated with a prey strain containing an empty vector (following mating steps of Subheading 3.4.2). 2. Cut a nitrocellulose membrane to the dimensions of an Omnitray plate (Nunc). Place the nitrocellulose membrane on top of diploid yeast colonies and leave for 10 s. 3. Use tweezers to lift the filter and slowly submerge in liquid nitrogen for 1 min. 4. Place the membrane on an empty Omnitray plate (Nunc) to thaw. 5. Cut a Whatman paper to same size as nitrocellulose membrane. Soak the Whatman paper with 2 ml Z-buffer, to which 35 ml X-solution had been added. 6. Overlay the nitrocellulose filter with the Whatman paper and remove air bubbles. 7. Incubate at 30°C for 10–60 min. 8. Evaluate: A blue stain indicates the activation of the betagalactosidase reporter and, therefore, a positive interaction.

4. Notes 1. Bait and prey must be transformed into yeast strains of opposite mating types to combine bait and prey plasmids by mating and to co-express the fusion proteins in diploids. Bait and prey plasmids can go into either mating type. However, this decision also depends on existing bait or prey libraries to which the new library may be mated later. 2. Baits are first grown at the different positions of a 96-well plate as liquid culture, and then cells are transferred (manually or with the use of a robot) to solid agar single-well plates (Omnitray plates). In this step, the 96-well format can also be converted into the 384-well format. This will position each bait in quadruplicates on the 384-well formatted plate. Full media agar (YEPD

Analysis of Protein–Protein Interactions

243

agar) can be used; however, for long-term storage of the array, selective agar is suggested to prevent loss of plasmids. 3. In a systematic array-based Y2H screening, duplicate or quadruplicate prey arrays are usually used. In a random genomic library screening, the entire experiment should be done in two copies to ensure reproducibility. Ideally, the template prey array should be kept on selective plates. The template of the prey array should be used to make “working” copies on YEPD agar plates for mating. The template can be used for 1–2 weeks; after 2 weeks it is recommended to copy the array onto fresh selective plates. Preys or bait clones tend to lose the plasmid if stored on YEPD for longer periods, which may reduce the mating and screening efficiency. 4. If the bait strains are frozen, they are streaked or pinned on selective solid medium plates and grown for 1–2 days at 30°C. Baits from this plate are then used to inoculate the liquid YEPD medium. It is important to make a fresh bait culture for Y2H mating, as keeping the bait culture on rich medium (YEPD) for a long time may cause loss of plasmids. Usually we grow baits overnight for mating. 5. Mating will usually take place in